AI-POWERED SMART SHOPPING: SEAMLESS

PERSONALIZATION

A PROJECT REPORT
Submitted by

NIKITHA S (111521203036)
MANVITHA R (111521203040)
RUPIKA N (111521203042)
VAISHNAVI R (111521203055)

in partial fulfillment for the award of the degree

of
BACHELOR OF TECHNOLOGY
in
INFORMATION TECHNOLOGY
R.M.D. ENGINEERING COLLEGE, THIRUVALLUR
(An Autonomous Institution)

MARCH 2025
 BONAFIDE CERTIFICATE

Certified that this project report titled “AI-POWERED SMART SHOPPING :

SEAMLESS PERSONALIZATION”, is the bonafide work of “NIKITHA S
(111521203036), MANVITHA R (111521203040), RUPIKA N (111521203042),
VAISHNAVI R (111521203055)” who carried out the project work under my
supervision, for the partial fulfillment of the requirements for the award of the degree
of Bachelor of Technology in Information Technology. Certified further that to the best
of my knowledge and belief, the work reported herein does not form part of any other
thesis or dissertation on the basis of which a degree or an award was conferred on an
earlier occasion.

SIGNATURE SIGNATURE

Dr. K. Balasubadra M.E., Ph.D., Dr. K. Balasubadra M.E., Ph.D.,

HEAD OF THE DEPARTMENT SUPERVISOR

Professor, Professor and Head,
Dept of Information Technology, Dept of Information Technology,
R.M.D. Engineering College, R.M.D. Engineering College,
R.S.M. Nagar, R.S.M. Nagar,
Kavaraipettai-601206. Kavaraipettai-601206.

ii
 CERTIFICATE OF EVALUATION

College Name : R.M.D. ENGINEERING COLLEGE

Department : INFORMATION TECHNOLOGY

Semester : 08

Name of the Students Name of the Supervisor

Title of Project with Register with Designation
Numbers

NIKITHA S

(111521203036)

MANVITHA R

AI-POWERED (111521203040) Dr. K. Balasubadra

SMART SHOPPING: M.E., Ph.D.,
SEAMLESS RUPIKA N
PERSONALIZATION Professor
(111521203042)

VAISHNAVI R

(111521203055)

The report of the project work submitted by the above students in partial fulfillment for
the award of Bachelor of Technology Degree in INFORMATION TECHNOLOGY of
Anna University was evaluated and confirmed to be the report of the work done by the
above students and then evaluated.

Submitted the project during the viva voce held on ...………..

INTERNAL EXAMINER EXTERNAL EXAMINER

iii
 ACKNOWLEDGEMENT

A project of this magnitude and nature requires the kind co-operation and support from
many, for the successful completion. We wish to express our sincere thanks to all those
who were involved in the completion of this project.
It is our immense pleasure to express our deep sense of gratitude to our Chairman
[Link], Vice-Chairman [Link], and our
Director [Link] NAIDU, for the facilities and support given by them in the
college.
We are extremely thankful to our Principal [Link], for having given
us an opportunity to serve the purpose of any education.
We wish to express our special thanks to our Dean-Research
[Link] for their constant support and encouragement.
We are indebted to [Link], M.E, Ph.D., Professor, Head of the
Department of Information Technology, for her valuable guidance and useful
suggestions during the course of the project.
We are thankful to our project supervisor Dr. K. BALASUBADRA, M.E, Ph.D.,
Professor and Head in the Department of Information Technology, R.M.D. Engineering
College for her guidance and valuable support given to us throughout the project.

iv
 ABSTRACT

The Intelligent Retail Experience revolutionizes traditional shopping by integrating

advanced technologies such as facial recognition, voice assistance, and automated payment
systems. Customers begin by registering their personal information, including email, phone
number, and facial data, to enable a seamless and personalized shopping experience.
Returning customers benefit from quick information retrieval, allowing easy store access.
Once inside, a voice assistant helps navigate the store, locate products, and provide real-
time pricing based on quantity and weight. As items are selected, the system updates the
customer’s account with purchase details, ensuring a smooth shopping process.

Upon completing their shopping, customers receive an electronic bill and a QR code via
email for convenient payment. An additional email verification step enhances security
during store entry. This intelligent system streamlines the entire shopping journey,
eliminating checkout lines and offering a personalized experience through automation. By
integrating convenience, security, and efficiency, the Intelligent Retail Experience
transforms the way customers interact with retail stores, ensuring a frictionless transition
from product selection to payment.

The Intelligent Retail Experience revolutionizes shopping by integrating automation,

security, and seamless transactions. Upon entering the store, customers verify their email
for secure access. As they shop, smart sensors track selected items, eliminating the need for
manual checkout. Once shopping is complete, an electronic bill and a QR code for instant
payment are sent via email, ensuring a hassle-free process.

AI-driven personalization tailors recommendations based on purchase history, enhancing

the overall experience. This system reduces wait times, minimizes errors, and enhances
customer satisfaction by offering a frictionless journey from product selection to payment,
redefining modern retail convenience.

v
 TABLE OF CONTENTS

CHAPTER TITLE PAGENO

ABSTRACT V
LIST OF FIGURES IX
LIST OF ABBREVIATIONS X
1 INTRODUCTION
1.1 GENERAL 1

1.2 AI-POWERED SMART SHOPPING 1

1.3 OBJECTIVE 2
1.4 EXISTING SYSTEM
1.4.1 Challenges 3
1.4.2 Literature Survey 4

1.5 PROPOSED SYSTEM 6

1.5.1 Advantages of Proposed System 6
[Link] Enhanced Convenience 7
[Link] Real-Time Assistance 7
[Link] Automated Billing and Payment 8

2 PROJECT DESCRIPTION
2.1 GENERAL 9
2.2 SYSTEM CONFIGURATION 9

2.3 ARCHITECTURE DIAGRAM 10

2.4 LIST OF MODULES 10
2.4.1 User Registration Module 10
2.4.2 Voice Assistance and Payment Module 10
2.4.3 Facial Recognition Using RESNET 11
2.4.4 NLP Technology 11

2.5 PROPOSED SYSTEM ALGORITHM

2.5.1 Introduction to RESNET Algorithm 12
2.5.2 Residual Learning 12
2.5.3 Variants of RESNET 14
2.5.4 Applications of RESNET 15

vi
3 REQUIREMENT ENGINEERING
3.1 GENERAL 16
3.2 HARDWARE REQUIREMENTS 16
3.2.1 Server Infrastructure 16
3.2.2 Storage Devices 17
3.2.3 Processing Units 17
3.2.4 Networking Elements 17

3.3 SOFTWARE REQUIREMENTS 18

3.3.1 Database Management System 18
3.3.2 Python with Flask Framework 18
3.3.3 Security Measures 18

4 SYSTEM DESIGN
4.1 GENERAL 19
4.1.1 Class Diagram 19
4.1.2 Use-Case Diagram 21
4.1.3 Activity Diagram 23
4.1.4 Data Flow Diagram 25

4.2 ARCHITECTURAL DESIGN 28

4.2.1 Front-end Design 28
4.2.2 Back-end Design 29

DEVELOPMENT TOOLS
5.1 GENERAL 31
5 5.2 TECHNICAL DESCRIPTION 31
5.2.1 Databases 31
5.2.2 GUI’s 31
5.2.3 Edge Computing Devices 32
32
5.2.4 Material UI
32
5.2.5 Secure Communication 32
5.2.6 Access Control and Authentication
5.3 BACKEND DEVELOPMENT TOOLS
5.3.1 Flask 33
5.3.2 Python 33

vii
6
RESULTS AND DISCUSSION
6.1 GENERAL 34
6.2 PROCEDURES AND SCREENSHOTS 34

6.3 EVALUATION METRICS 51

6.3.1 Quantitative Metrics 51
6.3.2 Qualitative Metrics 53
7 TESTING
7.1 GENERAL 54
7.1.1 Strategic Approach to Software Testing 54
7.2 UNIT TESTING 55
7.3 BASIC PATH TESTING 55
7.4 INTEGRATION TESTING 56
7.5 DATA FLOW TESTING 56
7.6 USER ACCEPTANCE TESTING 57
7.7 LOOP TESTING 58

8 CONCLUSION AND FUTURE ENHANCEMENTS

8.1 CONCLUSION 59
8.2 FUTURE ENHANCEMENTS 60

REFERENCES 61

viii
 LIST OF FIGURES

FIGURE NO. TITLE PAGE

NO.
2.3 ARCHITECTURE DIAGRAM 10

[Link] CLASS DIAGRAM OF RELATIONSHIP WITH 20

EACH CLASSES
[Link] USE CASE DIAGRAM FOR INTELLIGENT 22
RETAIL EXPERIENCE SYSTEM
[Link] ACTIVITY DIAGRAM FOR INTELLIGENT 25
RETAIL EXPERIENCE SYSTEM
[Link] DATA FLOW DIAGRAM FOR INTELLIGENT 27
RETAIL EXPERIENCE SYSTEM
6.2.1 HOME PAGE 47

6.2.2 REGISTRATION PAGE 48

6.2.3 SUPERMARKET DASHBOARD 49

6.2.4 SUPERMARKET BILL 50

ix
 LIST OF ABBREVATIONS

AI - ARTIFICIAL INTELLIGENCE

NLP - NATURAL LANGUAGE PROCESSING

ML - MACHINE LEARNING

UAT - USER ACCEPTANCE TESTING

POS - POINT OF SALE

API - APPLICATION PROGRAMMING INTERFACE

GUI - GRAPHICAL USER INTERFACE

SDK - SOFTWARE DEVELOPMENT KIT

DFD - DATA FLOW DIAGRAM

DBMS - DATABASE MANAGEMENT SYSTEM

P2P - PEER TO PEER

KYC - KNOW YOUR CUSTOMER

IDE - INTEGRATED DEVELOPMENT ENVIRONMENT

UI - USER INTERFACE

UX - USER EXPERIENCE

OTP - ONE-TIME-PASSWORD

x
 CHAPTER 1

INTRODUCTION

1.1 GENERAL

In today's technology-driven world, the retail industry is undergoing a profound

transformation with the integration of intelligent systems. Traditional shopping methods,
while familiar, often struggle with inefficiencies, long checkout queues, and limited
personalization. As consumer expectations evolve, there is an increasing demand for
seamless, secure, and efficient shopping experiences. This shift has paved the way for the
Intelligent Retail Experience, a concept that leverages advanced technologies such as facial
recognition, voice assistance, and automated payment systems to redefine how customers
interact with retail spaces.

By integrating these cutting-edge innovations, the Intelligent Retail Experience aims to

eliminate friction points, enhance security, and provide a highly personalized journey for
shoppers. From effortless store entry using facial recognition to real-time product assistance
and automated billing, this approach streamlines the entire retail process. The transition
from conventional methods to intelligent systems not only improves efficiency but also
enhances customer satisfaction. As digital transformation reshapes consumer behavior, the
adoption of such smart retail solutions marks a significant step towards the future of
shopping, offering unparalleled convenience and redefining the retail landscape.

1.2 AI-POWERED SMART SHOPPING

In today’s fast-paced world, customers demand convenience, personalization, and

efficiency in every aspect of their shopping experience. Traditional retail systems often
fall short in meeting these expectations, leaving room for improvement in customer
satisfaction and store operations. The Intelligent Retail Experience concept addresses
these needs by integrating advanced technologies such as facial recognition,
voice assistance, and automated billing into the shopping process. This innovative system

1
 ensures that customers receive a seamless, efficient, and personalized shopping
experience from the moment they step into the store. With this system, customers no
longer need to manually check out or search for products. By simply registering their
details, including email, phone number, and facial data, they can access their personalized
shopping profile upon entry. Returning customers benefit from swift data retrieval,
eliminating the need to input details again. Inside the store, the voice assistant acts as a
helpful guide, providing real-time product information, prices based on quantity and
weight, and recommendations based on previous purchases or preferences.
This instant access to information enhances decision-making and makes the shopping
process smoother and faster.

1.3 OBJECTIVE
The primary objective of the Intelligent Retail Experience is to leverage advanced
technologies like facial recognition and voice assistance to streamline the shopping
process, benefiting both customers and store owners. By utilizing a ResNet-based model
for facial recognition, the system ensures a quick and secure registration process for
customers, allowing for personalized services and easy access to their shopping profiles.
This facial recognition feature enhances security and convenience, as returning customers
can enter the store without the need for manual check-ins, ensuring a seamless experience.
The integration of voice assistance powered by Natural Language Processing (NLP)
further enhances the shopping journey. Customers can interact with the system by simply
asking about product availability, pricing, and other relevant information, making the
shopping process more intuitive and efficient. For sellers, these features simplify store
management by automating customer interactions, product inquiries, and real-time
pricing updates, which reduces manual effort and improves operational efficiency.

2
 1.4 EXISTING SYSTEM

1.4.1 CHALLENGES

The existing systems in retail environments typically rely on traditional methods for
customer registration, product discovery, and payment processing. Customers usually
register manually at the entrance, requiring them to fill out forms or provide basic
information such as name and contact details. Product information is often accessed via in-
store signage or by asking store attendants, leading to delays and a lack of real-time updates
on product availability or pricing.

Additionally, checkout processes are primarily manual, involving queues and cashier
interactions, which can be time-consuming and cumbersome for both customers and store
staff. In terms of security and personalization, many stores still rely on basic loyalty
programs or membership cards, which require customers to remember or carry physical
information. These systems do not integrate facial recognition or voice assistance, meaning
that the store experience lacks the level of automation, convenience, and efficiency that
modern consumers increasingly expect.

Customers need to ask store attendants or rely on physical signage for product
information, which may not be updated in real-time. Traditional checkout systems often
involve long queues and manual processing, causing delays and customer dissatisfaction.
Customers may not be aware of product availability until they reach the checkout or ask a
store attendant, leading to frustration if the item is out of stock.

3
1.4.2 LITERATURE SURVEY

PAPER 1:
TITLE: Radar Self-Following Shopping Cart Based on Multi-Sensor Fusion AUTHOR:
Lei yao, Tianhao Li, Rui Tong, Kai Wang, and Lingxiang Zhang YEAR: 2023

ABSTRACT:
Long queues of people waiting for a settlement in global shopping supermarkets are
bustling with traffic. In order to further improve operating efficiency and reduce the cost
of manual operation of the supermarket, this article proposes a multi-sensor integration
shopping cart system, which mainly includes image recognition, radar self-following
technology and road patrol. The improved SIFT algorithm effectively identifies a variety
of commodities. At the same time, relying on the public cloud platform, discount
information is recommended through WeChat programs, and the development of member
users login interface. After comprehensive calculation, the total price is displayed by the
EAIDK-610 user interface edited by Tkinter to complete the checkout. The LD-14 radar
used can realize the functions of autonomous follow-up and pedestrian obstacles. When
a sensor in the system fails or the detection deviation is too large, the multi-sensor fusion
technology can ensure the normal operation of the system due to a certain degree of
redundant. Improve the credibility and effectiveness of the data as a whole. This not only
solves the problem of queuing and checkout in the supermarket, but also helps brand
merchants to achieve precise marketing and scenario access. Finally, the use of computer
vision technology and various embedded hardware devices to build a radar self-following
the hardware prototype of the shopping cart for experimental verification. The average
accuracy of APs in the system is as high as 98.31%and 97.28%, respectively.

4
PAPER 2:

TITLE: A Hybrid Scheduling Mechanism Based on Agent Cooperation Mechanism and

Fair Emergency First in Smart Factory

AUTHOR: Sehrish Malik 1 and Dohyeun Kim YEAR: 2020

ABSTRACT:

Smart factory also known as smart manufacturing is an emerging field with the
revolution of industry 4.0. With the help of all these concepts, the smart factory integrates
the manufacturing assets and represents industrial networks. In this paper, we focus on
integrated solutions for smart factory concerns; by proposing an efficient task management
mechanism based on an efficient and resource-aware scheduling scheme named as ACM-
FEF. The scheduling algorithm used for the efficient task management is hybrid of the two
scheduling approaches as agent cooperation mechanism (ACM) and fair emergency first
(FEF) scheduling scheme. ACM is a decentralized scheduling approach which focuses on
the production maximization goals per machine, and also pays attention to the production
goals of all the machine networks involved in the smart factory. FEF scheduling scheme
focuses on minimizing the tasks starvation rate and maximizing the machine utilization by
efficiently using the machine slots. The proposed hybrid mechanism aims to efficiently plan
tasks execution, maximize machines’ resource utilization, maximize productivity,
minimize production delays, efficiently handle exceptions and efficiently control smart
factory Actuators

5
1.5 PROPOSED SYSTEM

The proposed system, the Intelligent Retail Experience, aims to revolutionize the
shopping process by integrating advanced technologies such as facial recognition, voice
assistance, and automated billing. Upon entering the store, customers will be prompted to
register their details, including email, phone number, and facial data. This registration
process ensures that both new and returning customers can quickly access their
personalized shopping profiles. For returning customers, the system will automatically
retrieve their information using facial recognition, streamlining entry and improving
security. Inside the store, the voice assistant, powered by Natural Language Processing
(NLP), will allow customers to inquire about product availability, pricing, and
specifications.
Customers can interact with the system using voice commands, and the assistant will
provide real-time responses, ensuring a smoother and more efficient shopping experience.
If a product is unavailable, the voice assistant will immediately inform the customer,
saving time and preventing frustration. Once the customer selects the desired products,
the system will automatically update their account with purchase details, including the
product’s weight and price. At checkout, there will be no need for manual processing, as
the system generates an electronic bill and a QR code, which is sent directly to the
customer’s email. This allows for seamless payment via the QR code, reducing wait times
and enhancing convenience. This project focuses solely on the software development
aspect, serving as a prototype for the Intelligent Retail Experience without involving any
hardware components.

6
 1.5.1 ADVANTAGES OF THE PROPOSED SYSTEM

In this pivotal sub-section, the Intelligent Retail Experience is introduced,

highlighting a range of advantages that address the inefficiencies in traditional retail
shopping. The proposed system is meticulously designed to leverage cutting-edge
technologies such as facial recognition, voice assistance, and automated billing, ensuring a
transformative shift in the retail landscape.

[Link] ENHANCED CONVENIENCE

One of the primary advantages of the proposed system is its ability to offer a
seamless and highly convenient shopping experience. By integrating facial recognition
for store entry and automated billing for purchases, the system eliminates the need for
manual authentication and checkout. This hands-free approach minimizes friction in the
shopping process, allowing customers to shop efficiently without unnecessary delays.

[Link] REAL-TIME ASSISTANCE

The integration of a voice assistant powered by Natural Language Processing (NLP)

enhances real-time customer support. Shoppers can instantly inquire about product
availability, pricing, and specifications using voice commands, receiving immediate and
accurate responses. This not only improves decision-making efficiency but also eliminates
the need for customers to manually search for information, significantly enhancing their
overall shopping experience.

7
 [Link] AUTOMATED BILLING AND PAYMENT
Efficiency is a key highlight of the proposed system. The automated billing process
eliminates long queues and manual checkout, allowing customers to complete their
purchases effortlessly. A QR code is generated and sent via email, enabling seamless
digital payment and reducing wait times. This automation not only enhances customer
satisfaction but also optimizes store operations by reducing the dependency on cashiers
and manual processing.

[Link] SECURE AND PERSONALIZED EXPERIENCE

By leveraging facial recognition for user authentication and personalized account

retrieval, the system ensures a secure and tailored shopping experience. Returning
customers benefit from quick access to their profiles, while security measures such as
email verification further enhance protection against unauthorized access. This
personalized approach fosters customer engagement and trust, reinforcing a future-ready
retail model that adapts to evolving consumer expectations.

[Link] IMPROVED OPERATIONAL EFFICIENCY

The proposed system optimizes overall store operations by reducing human

intervention in checkout, customer support, and security processes. With automated entry,
real-time product assistance, and self-checkout features, retailers can allocate resources
more efficiently, reducing labor costs while enhancing service quality. This streamlined
approach enables a more productive and cost-effective retail environment, benefiting both
businesses and customers alike.

8
 CHAPTER 2

PROJECT DESCRIPTION

2.1 GENERAL
This section provides a broad yet comprehensive overview of the project, outlining
its key components, technologies, and methodologies within the context of intelligent
retail shopping. The Intelligent Retail Experience leverages facial recognition for
seamless authentication, integrates voice assistance powered by NLP for real-time
customer support, and implements automated billing to streamline the checkout process.
The methodologies involve software-based smart retail system development, data
processing for personalized shopping, and secure digital transactions through QR-based
payments. The significance of this project lies in its ability to redefine the retail shopping
experience by enhancing convenience, security, and operational efficiency, paving the
way for the future of smart retail.

2.2 SYSTEM CONFIGURATION:

H/W SYSTEMCONFIGURATION:
• Processor – Intel
• Speed - 1.1 GHz
• RAM - 8 Gb
• Hard Disk - 500 GB
S/W SYSTEM CONFIGURATION:
• Operating System - Windows 7/8/10

• Front End - HTML,CSS

• Scripts - Python Language
• Tool - Python IDLE

9
2.3 ARCHITECTURE DIAGRAM

2.4 LIST OF MODULES

• User Registration Module

• Voice Assistance and Payment Module
• Facial Recognition using Resnet Module
• Natural Language processing Technology

MODULE DESCRIPTION:

2.4.1 USER REGISTRATION MODULE:

The User Registration Module is responsible for registering customers when
they enter the store. This module integrates both facial recognition and email
registration into a single, seamless process. New customers will provide their
email addresses and facial data, which are stored securely in the system.
Returning customers will have their profiles automatically retrieved using facial
recognition technology, ensuring fast and secure entry.

2.4.2 VOICE ASSISTANCE AND PAYMENT MODULE:

The Voice Assistance and Payment Module enhances the shopping experience

10
 by providing customers with an intuitive, hands-free interface. Using Natural
Language Processing (NLP) techniques, the voice assistant helps customers
inquire about product availability, pricing, and specifications, making their
interaction with the system more efficient. Additionally, this module handles
payment processing by automatically generating a digital bill and sending a QR
code to the customer's email for seamless payment. By integrating voice
assistance and payment functionality, this module simplifies the purchasing
process and minimizes wait times at checkout.

2.4.3 FACIAL RECOGNITION USING RESNET MODULE:

The Facial Recognition Using ResNet Module is designed to identify customer
based on their facial features. Using the ResNet deep learning model, the system
accurately processes facial data to recognize returning customers and retrieve
their personalized information. This ensures a quick and secure customer entry
into the store, eliminating the need for manual check-ins. The module improves
both security and customer experience,making the entire shopping journey more
personalized and efficient by automatically linking
the customer to their profile upon entry.

2.4.4 NATURAL LANGUAGE PROCESSING TECHNOLOGY:

Voice Assistance Using NLP Module provides an interactive, voice-driven
interface for customers to query the system about various aspects of the store,
such as product availability, prices, and specifications. Powered by advanced NLP
techniques, the module ensures that the system accurately understands and
responds to voice commands . It streamlines product discovery and enhances the
customer’s shopping experience by allowing them to browse and gather
information without having to rely on physical [Link] module serves as a key
interface for customers to interact with the system using voice commands.

11
2.5 PROPOSED SYSTEM ALGORITHM:
2.5.1 INTRODUCTION TO RESNET ALGORITHM
ResNet (Residual Network) is a powerful and widely adopted deep learning
architecture, first introduced in the paper “Deep Residual Learning for Image Recognition”
by Kaiming He et al. in 2015. The primary aim of ResNet is to solve the problem of training
very deep neural networks. As networks deepen, they tend to suffer from performance
degradation due to vanishing gradients, poor convergence, and increased complexity, but
ResNet overcomes these challenges with its innovative architecture. It introduces the
concept of residual learning, which helps networks achieve much deeper layers while
maintaining efficient training.

ResNet has been a breakthrough in the field of computer vision, winning the 2015
ImageNet competition with outstanding performance. Since its introduction, ResNet has
become a foundational architecture for various applications, from image recognition to
more advanced tasks like object detection, segmentation, and even natural language
processing.

2.5.2 Core Idea: Residual Learning

Before ResNet, increasing the depth of neural networks was thought to improve their
capacity to model complex data. However, adding more layers to a deep network also
caused training difficulties. When networks become very deep, it is hard to learn effective
representations due to the vanishing gradients problem. This happens because gradients
used in backpropagation can become extremely small, making it difficult for weights in
earlier layers to change during training. ResNet tackles this problem by introducing skip
connections or residual connections, which are short paths that bypass one or more layers.
In a residual network, the output of each layer is added to the output of the previous layer,
creating what is known as a residual block.

The key idea of residual learning is expressed mathematically as: y=F(x,{Wi})+xy =

F(x,\{W_i\}) + xy=F(x,{Wi })+x where:

12
 F(x,{Wi})F(x, \{W_i\})F(x,{Wi }) is the output of the residual block
after applying a series of transformations (convolutions, activations, etc.),

yyy is the final output of the block.

Here, instead of learning the direct mapping y=F(x)y = F(x)y=F(x), the network learns
the residual function F(x)F(x)F(x) that is added to the input xxx. This makes it easier for
the network to learn identity mappings (i.e., when the input and output are the same), which
helps prevent the performance degradation associated with deep networks.

Architecture of ResNet:

The ResNet architecture consists of a stack of residual blocks, and there are different
variants of ResNet depending on the number of layers. Common variants are ResNet-18,
ResNet-34, ResNet-50, ResNet-101, and ResNet-152, where the numbers represent the total
number of layers in the network.

A standard residual block in ResNet typically consists of:

1. Two convolutional layers (often 3x3 filters),
2. Batch normalization after each convolution to stabilize learning,
3. ReLU activation functions (Rectified Linear Units),
4. Skip connections that bypass the convolutional layers and directly add the
input to the output.
The deeper ResNet versions, such as ResNet-50, ResNet-101, and ResNet-152, use more
sophisticated residual blocks with bottleneck structures. A bottleneck residual block has
three layers:
• A 1x1 convolution to reduce dimensionality,
• A 3x3 convolution for the core processing,
• A 1x1 convolution to restore dimensionality.

Why ResNet Works

The effectiveness of ResNet comes from several factors:
• Avoiding vanishing gradients: The introduction of skip connections allows the
gradients to flow more easily through the network during backpropagation.
13
 • This ensures that gradients remain large enough for earlier layers to update their
weights effectively.
• Residual learning is easier: Instead of learning a direct mapping, the network learns
the residual, which is easier to optimize. This leads to better performance, especially
as the network depth increases.
• Improved convergence: Deep residual networks are much easier to train compared
to traditional deep networks. ResNet facilitates faster convergence and reduces
training time.
• Deeper networks generalize better: With residual connections, the network can
reach much deeper architectures without overfitting. This allows the model to
capture more complex patterns in the data.
• Building blocks for transfer learning: ResNets architecture has become a standard
backbone for transfer learning, where pretrained networks on large datasets like
ImageNet are fine-tuned for specific tasks. The deep network and efficient learning
make it an ideal candidate for feature extraction in other domains, such as medical
imaging or video analysis.

2.5.3 Variants of ResNet

• ResNet-18 and ResNet-34: These are smaller versions of ResNet, with fewer
layers. They are computationally less expensive and work well for relatively
simpler tasks. ResNet-18 and ResNet-34 use basic residual blocks without the
bottleneck structure.
• ResNet-50, ResNet-101, and ResNet-152: These are larger versions of ResNet,
using bottleneck residual blocks for greater efficiency in training. ResNet-50, for
instance, contains 50 layers and is commonly used for more complex tasks like
object detection or segmentation, where higher model capacity is needed.
• ResNetV2: A variant of ResNet called ResNetV2 was proposed in 2016 by He et al.
This version introduces the idea of pre-activation, where batch normalization and
ReLU are applied before the convolution operation. This improves performance and
helps train even deeper networks.
• Wide ResNet: Another variation, known as Wide ResNet, increases the number of
14
 filters (width) at each layer rather than the depth of the network. Wide ResNet
models are more efficient in terms of computational cost and have shown
competitive performance for certain tasks.

2.5.4 Applications of ResNet

ResNet has demonstrated state-of-the-art performance across many computer vision
tasks, including:
• Image classification: ResNet is widely used for classifying images into different
categories. Its performance on the ImageNet challenge has set a new benchmark for
image classification tasks.
• Object detection: By acting as a backbone for models like Faster R-CNN or Retina
Net, ResNet provides powerful feature and improving detection accuracy.
• Semantic segmentation: In tasks where it’s necessary to label each pixel of an
image (e.g., in autonomous driving), ResNet has been integrated into segmentation
frameworks like DeepLab.
• Face recognition: Deep residual networks have been successfully applied in face
recognition, where their ability to learn intricate features and patterns has proven
useful in identifying individuals.
• Medical imaging: In domains like radiology, ResNet-based models are trained to
detect tumors, lesions, and other abnormalities in medical images, aiding in
diagnosis and treatment planning.
• Generative models: Although ResNet is predominantly used for discriminative
tasks, it has also been employed in generative models like GANs (Generative
Adversarial Networks) for generating high-quality images.
The ResNet architecture has made a transformative impact on deep learning. Its
innovative use of residual connections has enabled the training of ultra-deep neural
networks without facing the issues of vanishing gradients or overfitting. As a result, ResNet
has achieved state- of-the-art results across a wide range of computer vision tasks. Its
versatility has made it the go-to model for many applications, from classification to
segmentation and beyond.

15
 CHAPTER 3

REQUIREMENT ENGINEERING

3.1 GENERAL

In this foundational section of Requirement Engineering, we outline the key

considerations and principles guiding the development of the Intelligent Retail
Experience. Requirement engineering serves as the framework for ensuring that both
hardware and software components align with the objectives of the project. This process
involves collaboration with stakeholders, understanding user needs, and establishing a
structured approach to requirement elicitation and analysis. The proposed system focuses
on a software-based prototype that integrates facial recognition, voice assistance, and
automated billing to enhance the retail shopping experience.

3.2 HARDWARE REQUIREMENTS

This section provides a detailed overview of the necessary hardware components

essential for the optimal performance of the Intelligent Retail Experience. Each hardware
element is carefully chosen to support the system’s functionality, ensuring efficiency,
reliability, and scalability.

3.2.1 SERVER INFRASTRUCTURE

The server infrastructure plays a critical role in the Intelligent Retail Experience,
ensuring efficient data processing, real-time interactions, and secure transactions. A
cloud-based or dedicated server is required to handle customer authentication, voice
assistant queries, and automated billing. The server must support high-speed data retrieval
and concurrent request handling, enabling smooth customer interactions and real-time
updates. Additionally, a scalable server architecture is essential to accommodate growing
user demands and ensure seamless system performance.

16
3.2.2 STORAGE DEVICES

The selection of storage devices is a crucial aspect of the hardware requirements,

driven by the need for reliable and scalable solutions. The system demands high-
performance storage devices, such as SSDs or HDDs, to support a growing database and to
handle increasing amounts of customer interactions and transactions over time.
Additionally, cloud storage integration can be utilized for real-time data synchronization,
backup, and secure access, ensuring seamless system operation with minimal risk of data
loss.

3.2.3 PROCESSING UNITS

The Intelligent Retail Experience relies on a powerful multi-core processor to
efficiently handle facial recognition, voice assistance, and automated billing operations.
An Intel processor with a speed of 1.1 GHz is selected to support real-time data processing,
ensuring smooth execution of complex tasks such as image recognition, natural language
processing (NLP), and transaction updates. Given the computational demands of AI-driven
features, the processor must provide high-speed performance, multitasking capabilities,
and energy efficiency to maintain system responsiveness. A well-optimized processing
unit ensures that customer interactions, product lookups, and billing processes are executed
seamlessly, delivering a fast and reliable shopping experience.

3.2.4 NETWORKING ELEMENTS

Efficient system communication is enabled through a high-speed network interface,
ensuring uninterrupted data exchange between system components. A stable and reliable
Wi- Fi or Ethernet connection is required to support real-time cloud-based interactions for
customer authentication, product lookups, and automated billing.

Each hardware requirement is meticulously selected to optimize system performance,

ensure scalability, and enhance the overall user experience, forming a robust foundation
for the proposed system.

17
3.3 SOFTWARE REQUIREMENTS
The Intelligent Retail Experience is designed with specific software requirements
to ensure efficient operation, secure transactions, and an interactive shopping
experience.

3.3.1 MYSQL DATABASE MANAGEMENT SYSTEM

The system relies on MySQL, a robust relational database management system.
MySQL is chosen for its reliability, scalability, and ability to efficiently handle structured
data, providing a secure and organized repository for verified documents.

3.3.2 PYTHON WITH FLASK FRAMEWORK

The backend of the system is developed using Python as the programming language
and the Flask framework. Python's versatility and readability, combined with Flask's
simplicity and flexibility, create a powerful backend infrastructure. Flask facilitates the
rapid development of secure and scalable web applications, tailored to the requirements
of the rapid development, real-time data processing, and seamless integration with various
system modules such as voice assistance and automated billing.

3.3.3 SECURITY MEASURES

To enhance system security, industry-standard protocols are implemented:

• Secure authentication using facial recognition and email verification to

prevent unauthorized access.
• Encryption mechanisms for sensitive information such as customer details and
payment credentials.
• HTTPS implementation ensures encrypted communication between the user and
the system, securing data transmission.
• Regular security audits and updates to address potential threats and enhance
system resilience.

18
 CHAPTER 4

SYSTEM DESIGN

4.1 GENERAL
In the system design phase, the overall architecture and structure of the Intelligent
Retail Experience system are conceptualized and outlined. This phase lays the foundation
for the subsequent detailed designs and development. The key components of the general
system design include:

4.1.1 CLASS DIAGRAM

The Class Diagram serves as a foundational element in the system design of the
Intelligent Retail Experience. It provides a structured visual representation of the key
classes and their relationships, offering a clear understanding of how different system
components interact.

● Classes and Relationships:

o Customer Class: Represents shoppers who interact with the system. It

includes attributes such as customer ID, name, email, phone number, and
facial recognition data. This class manages user authentication and shopping
profiles

o Product Class: Represents the products available in the store. Attributes

include product ID, name, category, price, weight, and stock availability. It
connects with the voice assistant for product inquiries and updates the
customer's cart upon selection.

o Voice Assistant Class: Handles customer interactions using Natural

Language Processing (NLP). Attributes include query type, response data,
and product recommendations. It assists in product searches, pricing
inquiries, and general shopping guidance.

19
● Class Relationships:
• Customer-Product Relationship: Represents the interaction where
customer browse and select products. When a product is selected, it is linked
to the customer's shopping cart.
• Customer-Voice Assistant Relationship: Shows how customers engage with
the
voice assistant to inquire about product details, availability, and
recommendations.
• Customer-Billing Relationship: Demonstrates the connection between
customers and the billing system, which updates the purchase details and
generates an invoice.
• Product-Store Database Relationship: Establishes how product data is
stored and retrieved from the store database, ensuring accurate inventory
management.
• Billing-Store Database Relationship: Defines how transaction details are
recorded in the store database for future reference and customer convenience.

Figure.[Link] CLASS DIAGRAM OF RELATIONSHIP WITH EACH CLASSES

20
4.1.2 USE CASE DIAGRAM

The Use Case Diagram plays a crucial role in designing the Intelligent Retail
Experience system by defining the interactions between users and the system. It
highlights the functionalities available to different user roles, offering a high-level
perspective on how users engage with the system. This visual representation ensures
clarity regarding system behavior, providing stakeholders with an intuitive
understanding of user interactions. By categorizing distinct use cases, the diagram
serves as an essential tool for designing a user-centric and efficient shopping
experience.

● Actors:
1. Customer – Represents shoppers who interact with the system for product

selection, voice assistance, and automated checkout.

2. Administrator – Manages the overall system, including user registrations,
product inventory, and system configurations.
3. Retail Store Database – Stores customer details, product information, billing
records, and authentication logs.
● Use Cases:
1. Register/Login – Customers must register their details, including facial
recognition data, email, and phone number, to access the store.
2. Facial Recognition Entry – Returning customers can enter the store using facial
recognition for seamless access.
3. Product Selection–Customers browse and select products, update their virtual
cart.
4. Automated Billing – The system generates an electronic bill based on selected
products, weight, and quantity.
5. QR Code Payment – A QR code is sent to the customer’s email for fast and secure
payment.
6. System Management – Administrators oversee customer data, inventory update.
21
Figure.[Link] USE CASE DIAGRAM FOR INTELLIGENT RETAIL EXPERIENCE SYSTEM

22
4.1.3 ACTIVITY DIAGRAM
The Activity Diagram is a crucial component of the system design for the Intelligent
Retail Experience, visually representing the workflow of different system activities.
It illustrates the step-by-step execution of various actions in the system, ensuring a
clear understanding of the process flow from customer entry to payment completion.
This diagram aids in optimizing system functionality and improving efficiency.

● Actors:

1. Customer – Interacts with the system for shopping, voice assistance, and billing.
2. Voice Assistant –Provides real-time product details and helps customers navigate.
3. Retail System – Manages product selection, billing, and checkout.
4. Admin – Handles system operations, inventory updates, and user registration.
5. Database – Stores user credentials, product data, and transaction records.

● Activity Flow:
1. Customer Entry & Registration

• New customers register with their personal details (email, phone, facial data).
• Returning customers use facial recognition for authentication and seamless
access.

2. Product Search & Inquiry via Voice Assistant

• Customers ask the voice assistant about product availability, pricing, and
specifications.
• The assistant fetches details from the retail system and provides real-time
responses.

3. Product Selection & Virtual Cart Update

• The customer selects desired products, and the retail system updates the virtual
cart.
• The database records the selected items along with weight and price details.

23
4. Automated Billing & Invoice Generation
• The retail system generates an electronic bill once the customer finalizes
selections.
• The system sends an invoice with a QR code to the customer's email for
payment.

5. QR Code Payment & Transaction Completion

• The customer scans the QR code and makes a digital payment.

• The retail system verifies the transaction and updates the database.
• If payment is successful, the system confirms the purchase and logs transaction
history.

6. Exit & System Update

• Customers receive a purchase confirmation and exit the store.
• The admin updates product inventory and manages system performance.

This Activity Diagram effectively showcases the streamlined and automated

workflow of the Intelligent Retail Experience, ensuring a seamless, efficient, and
user-friendly shopping process.

24
 Figure.[Link] ACTIVITY DIAGRAM FOR INTELLIGENT RETAIL EXPERIENCE SYSTEM

4.1.4 DATA FLOW DIAGRAM

The Data Flow Diagram (DFD) provides a structured representation of how data
moves through the Intelligent Retail Experience system. It visually depicts input,
processing, and output interactions, ensuring clarity in system workflow. The DFD
comprises different levels that progressively break down the system into smaller
functional components.

25
● Actors :
• Customer – Engages with the system for shopping, product search, and
payments.
• Admin – Manages system operations, user data, and inventory updates.
• Payment Gateway – Processes transactions after the customer scans the QR
code.
•
● DFD Levels:

Level 0: Context Diagram

• Represents a high-level view of the system.

• Shows interactions between Customer, Admin, Payment Gateway, and Retail
System.
• Demonstrates data flow between actors and the system, focusing on inputs and
outputs.

Level 1: High-Level Data Flow

Expands on Level 0 by breaking down major processes:

• Customer Registration & Authentication (Data from user → User Database)
• Product Search & Selection (Customer queries → Product Database)
• Billing & Payment Processing (Purchase confirmation → Payment Gateway)
• Inventory & System Updates (Stock updates → Product Database)

Level 2: Detailed Process Flow

Provides a deeper breakdown of each process:

• User Registration: Captures details, verifies users, and stores credentials.
• Voice Assistant Interaction: Retrieves product data and guides customers.
• Cart & Billing System: Adds items to cart, calculates price, and generates
invoices.
• QR Code Payment: Processes transactions and updates the database.
• Exit & Inventory Update: Confirms purchase and modifies stock levels.
26
 • Data Stores (Repositories)

• User Database – Stores customer credentials, purchase history, and

authentication details.
• Product Database – Contains product details, pricing, and inventory stock.
• Transaction Database – Records billing, payment status, and purchase logs.

Figure.[Link] DATA FLOW DIAGRAM FOR INTELLIGENT RETAIL EXPERIENCE SYSTEM

This Data Flow Diagram (DFD) effectively illustrates the structured movement of data
within the Intelligent Retail Experience, ensuring a smooth, efficient, and automated
shopping process.

27
4.2 ARCHITECTURAL DESIGN
In the context of the Intelligent Retail Experience, the Architectural Design outlines
the high-level structure and organization of the system. This design phase encompasses
both Frontend and Backend architectures, detailing how different components interact to
achieve the project's objectives of automated shopping, voice assistance, and seamless
billing.

4.2.1 FRONT-END DESIGN

The Frontend Design of the Intelligent Retail Experience focuses on ensuring a

smooth and intuitive shopping journey for customers. It consists of key User Interface (UI)
and User Experience (UX) elements to facilitate navigation, product search, and payment
processing.

1. User Interface (UI):

• Customer Registration & Authentication Form: Enables new users to register and
existing customers to authenticate using facial recognition.
• Product Search & Inquiry Panel: Allows users to browse, filter, and search for
products using voice commands or text input.
• Shopping Cart Display: Shows selected items with price, weight, and quantity for
easy review before checkout.
• Billing & Payment Screen: Generates an electronic bill and QR code for hassle-free
digital payments.

2. User Experience (UX):

• Seamless Navigation: Ensures smooth transes smooth transitioning between different

sections like product search, cart, a product search, cart, and checkout.
• Real-Time Voice Assistance: Provides instant responses to product queries,
availability checks, and store navigation.
• Instant Feedback Mechanism: Notifies users of stock availability, purchase
confirmation, and payment success/failure.

28
3. Visual Elements:

• Live Shopping Assistance Interface: Displays product details when a customer

inquires using voice commands.
• Billing Summary Visualization: Presents a clear order breakdown before generating a
QR code for payment.
• Status Indicators: Shows real-time order progress, from product selection to
payment completion.

4.2.2 BACK-END DESIGN

The Backend Design of the system focuses on the server-side logic, database
management, and system security. It processes customer interactions, manages product
databases, and facilitates automated billing and payment transactions.
• Server Infrastructure:

o User & Product Databases: Stores customer profiles, purchase history, and
product details.
o Authentication & Security Modules: Handles facial recognition authentication
and safeguards customer data.
o Billing & Payment Processing System: Generates invoices and QR codes for
seamless checkout and payment verification.

• Natural Language Processing (NLP) Engine:

o Understands customer queries and provides relevant product information.

o Text-to-Speech & Speech Recognition: Enables real-time voice interaction
between the customer and the system.
o Product Information Fetching System: Retrieves and displays relevant product
details in response to customer queries.
• Security Measures:
o Facial Recognition for Authentication: Ensures secure, personalized shopping.
29
 o Data Encryption: Protects sensitive customer and transaction data using secure
communication protocols.
o Access Control Mechanisms: Restricts unauthorized access to customer
profiles, purchase records, and store inventory.

• Integration with Frontend:

o RESTful APIs: Facilitates communication between UI and backend components for

real-time updates and product retrieval.
o Standardized Data Exchange Formats: Uses JSON and XML to ensure efficient
data transfer across different components.
o Based Data Synchronization: Ensures that customer profiles, purchase history, and
product availability are updated in real-time across multiple devices, allowing
seamless transitions between shopping sessions.

This architectural design forms the foundation of the Intelligent Retail Experience,
ensuring a fast, secure, and automated shopping environment that enhances user satisfaction
and efficiency.

30
 CHAPTER 5

DEVELOPMENT TOOLS

5.1 GENERAL

In the developmental strides of the Intelligent Retail Experience Shopping Project,

selecting the right tools is crucial for creating a seamless, efficient, and innovative
system. These tools play a vital role in building a smart retail environment by integrating
AI-driven analytics, IoT-based tracking, and a responsive shopping interface. The
careful selection of these tools ensures smooth communication between the frontend,
backend, and smart retail hardware, delivering an enhanced shopping experience for
users.

5.2 TECHNICAL DESCRIPTION

5.2.1 DATABASES:
The total number of databases that were identified to build the system is [Link]
major part of the Databases is categorized as Administrative components and the user
components. The administrative components are useful is managing the actual master
data that may; be necessary to maintain the consistency of the system. The
administrative databases are purely used for the internal organizational needs and
necessities. The user components are designed to handle the transactional state that arise
upon the system whenever the general client makes a visit onto the system for the sake
of the report based information. The user components are scheduled to accept
parametrical information for the user as per the systems necessity.

5.2.2 GUI’s
For the flexibility of the user, the interface has been developed in graphical
user interface mode. The normal interface is applied through browser.
The GUI’s at the top level has been categorized as
1) Administrative user interface
2) Customer or general user interface

31
 The administrative user interface concentrates on the consistent information that is
practically, pact of the organizational activities and which needs proper authentication
for the data collection. The interfaces help the visitors with all the transactional states
like Data insertion, Data deletion and Data updating with the data search capabilities.
The general user interface helps the users upon the system in transactions through the
required services that are provided upon the system. The general user interface also
helps the ordinary user is managing their own information in a customized manner as
per their flexibilities.

5.2.3 EDGE COMPUTING DEVICES

Edge devices process data locally before sending it to the cloud, reducing latency
and improving response times. These are essential for ensuring fast customer checkout,
smart shelves, and dynamic pricing updates.

5.2.4 MATERIAL UI

Material UI provides pre-designed, customizable UI components, ensuring a

modern and intuitive shopping interface. Its responsive design principles allow the
application to function efficiently across mobile, tablet, and desktop devices.

5.2.5 SECURE COMMUNICATION

The system implements SSL/TLS protocols to encrypt data transmission, ensuring

customer information and transaction data remain protected.

5.2.6 ACCESS CONTROL AND AUTHENTICATION

Role-based access control (RBAC) is enforced, ensuring only authorized users

(administrators, store managers, and customers) can access specific features. OAuth and
JWT tokens secure login sessions and API requests.

32
5.3 BACKEND DEVELOPMENT TOOLS

5.3.1 FLASK
Flask, a lightweight and versatile Python web framework, plays a pivotal role in the
backend development of the Blockchain Document Verification Project. It facilitates the
creation of robust and scalable backend functionalities. Flask's simplicity and flexibility
make it an ideal choice for implementing backend logic, handling verification requests,
and ensuring smooth communication between the frontend and the backend. Leveraging
Flask in conjunction with development tools contributes to the overall efficiency and
effectiveness of the shopping experience.

5.3.2 PYTHON
Python is a high-level object-oriented programming language that was created by
Guido van Rossum. It is also called general-purpose programming language as it is used
in almost every domain we can think of.

The integration of Flask, a Python web framework, adds a layer of versatility to

the AI-Powered seamless Shopping project. While Flask is not explicitly mentioned in
the development tools section, its role in the backend cannot be understated. Python,
through Flask, contributes to the backend functionalities, handling various aspects of the
verification process. Flask's simplicity and flexibility make it an ideal choice for building
APIs and handling server-side logic. In this context, Flask is instrumental in managing
interactions between the frontend and the backend, ensuring seamless communication and
processing of verification requests. The incorporation of Python, with Flask as a
framework, showcases the project's adaptability and the utilization of diverse
technologies to achieve a robust and cohesive system.

Python’s standard library is very extensive, offering a wide range of facilities. The
library contains built-in modules (written in C) that provide access to system functionality
such as file I/O that would otherwise be inaccessible to Python programmers, as well as
modules written in Python that provide standardized solutions for many problems that
occur in everyday programming.

33
 CHAPTER 6
RESULTS AND DISCUSSION

6.1 GENERAL
This section presents the general outcomes and findings of the Intelligent
Retail Experience Shopping Project. It provides an overall evaluation of the system's
performance, user engagement, efficiency, and its ability to enhance the shopping
experience. The assessment highlights the system’s effectiveness in automating checkout,
personalizing shopping experiences, optimizing inventory management, and integrating
AI-driven analytics. This general overview sets the stage for a more detailed exploration
in subsequent sections, focusing on procedures, visual representations, and key evaluation
metrics. By offering a holistic perspective, this section helps stakeholders understand the
impact of this smart retail system in revolutionizing traditional shopping.

6.2 PROCEDURES AND SCREENSHOTS

This subsection outlines the specific procedures followed in the Intelligent Retail
Experience Shopping Project, supported by relevant screenshots. It provides a step-by-
step walkthrough of key processes, including:

• User Entry & Smart Recognition: Demonstrating how customers are recognized
upon entering the store via RFID sensors or facial recognition.
• Product Interaction & AI-Powered Suggestions: Showcasing the system’s ability
to track customer movements and suggest personalized offers.
• Automated Checkout & Payment Processing: Illustrating the seamless checkout
process where customers can walk out, and their purchases are automatically billed
via a digital wallet.
• Inventory Monitoring & Smart Shelf Management: Visualizing how IoT-
enabled smart shelves track product stock levels and update inventory in real-time.
• AI-Powered Analytics Dashboard: Presenting insights into customer behavior,
heatmaps, and sales trends, providing store managers with valuable decision-
making data.

34
6.2.1 Implementation

[Link] :
from flask import Flask, render_template, request, jsonify, redirect, url_for, session
import sqlite3
import os
import cv2
import numpy as np
import base64
import tempfile
import dlib
import random
import smtplib
from [Link] import MIMEMultipart
from [Link] import MIMEText
from [Link] import MIMEImage
import smtplib
from [Link] import formataddr
import qrcode
import smtplib
from [Link] import EmailMessage
from fpdf import FPDF
app = Flask(__name__)
app.secret_key = "your_secret_key"
DATABASE = '[Link]'
#manvithar32
conn = [Link](DATABASE)
cursor = [Link]()
[Link]('''

35
 CREATE TABLE IF NOT EXISTS users (
id INTEGER PRIMARY KEY AUTOINCREMENT,

name TEXT NOT NULL,

email TEXT NOT NULL,
phone TEXT NOT NULL,
image_path BLOB
)
''')
[Link]('''
CREATE TABLE IF NOT EXISTS orders (
id INTEGER PRIMARY KEY AUTOINCREMENT,
email TEXT,
product TEXT,
quantity REAL,
price REAL
)
''')
[Link]()
# Home page
@[Link]('/')
def index():
return render_template('[Link]')
# Registration Route
if not [Link]("face_images"):
[Link]("face_images")
app = Flask(__name__)
# Database setup
@[Link]('/')
def index():
return render_template('[Link]')

36
UPLOAD_FOLDER = 'static/images'
[Link](UPLOAD_FOLDER, exist_ok=True)
UPLOAD_FOLDERS = 'static/login'

[Link](UPLOAD_FOLDERS, exist_ok=True)
@[Link]('/register', methods=['POST'])
def register():
data = [Link]
name = data['name']
email = data['email']
phone = data['phone']
image_path = data['image_path'] # base64 image data
conn = [Link](DATABASE)
cursor = [Link]()
# Check if the email already exists
[Link]("SELECT * FROM users WHERE email = ?", (email,))
existing_user = [Link]()
if existing_user:
[Link]()
return jsonify({'status': 'error', 'message': 'Already registered'}), 400
# Decode the base64 string and save the image
image_data = base64.b64decode(image_path.split(',')[1]) # Strip metadata
image_folder = [Link](UPLOAD_FOLDER, email)
[Link](image_folder, exist_ok=True)
image_filename = f"{email}.png" # Create a unique filename
image_full_path = [Link](image_folder, image_filename)
with open(image_full_path, "wb") as image_file:
image_file.write(image_data)
# Store only the relative path in the database
relative_image_path = f"{UPLOAD_FOLDER}/{email}/{image_filename}"

37
 [Link]('''
INSERT INTO users (name, email, phone, image_path)
VALUES (?, ?, ?, ?)
''', (name, email, phone, relative_image_path))
[Link]()

[Link]()
return jsonify({'status': 'success', 'message': 'Registration successful', 'image_path':
relative_image_path})
def process_and_match_image(image_path):
training_data_folder = 'static/images' # Folder with pre-saved images
detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor("shape_predictor_68_face_landmarks.dat")
face_recognizer =
dlib.face_recognition_model_v1("dlib_face_recognition_resnet_model_v1.dat")
# Load the uploaded image
uploaded_image = [Link](image_path)
gray_uploaded = [Link](uploaded_image, cv2.COLOR_BGR2GRAY)
faces_uploaded = detector(gray_uploaded)
print(faces_uploaded)
if not faces_uploaded:
print('not work')
return None # No faces detected
# Extract face encoding from the uploaded image
for face in faces_uploaded:
shape_uploaded = predictor(gray_uploaded, face)
uploaded_face_encoding =
face_recognizer.compute_face_descriptor(uploaded_image, shape_uploaded)
#print('d', uploaded_face_encoding)
# Compare with images in the training data folder

38
 for person_name in [Link](training_data_folder):
print(person_name)
person_folder = [Link](training_data_folder, person_name)
print(person_folder)
if [Link](person_folder):
for filename in [Link](person_folder):
if [Link](('.jpg', '.jpeg', '.png')):
training_image_path = [Link](person_folder, filename)

training_image = [Link](training_image_path)
gray_training = [Link](training_image, cv2.COLOR_BGR2GRAY)
faces_training = detector(gray_training)
for face in faces_training:
shape_training = predictor(gray_training, face)
training_face_encoding =
face_recognizer.compute_face_descriptor(training_image, shape_training)
# Calculate distance between uploaded image and training image
distance = [Link]([Link](uploaded_face_encoding) -
[Link](training_face_encoding))
print(distance)
if distance < 0.5: # Threshold for a match
return person_name # Return the matched person's name
a=[]
# Login Route
@[Link]('/login', methods=['POST'])
def login():
data = [Link]
email = [Link]('email')
print("Entered Email:", email)
[Link](email)
image_path = [Link]('image_path')

39
 if not image_path:
return jsonify({'error': 'No image provided'}), 400
# Decode and save the image
try:
image_data = base64.b64decode(image_path.split(',')[1]) # Remove metadata
image_filename = f"{email}.png"
image_full_path = [Link](UPLOAD_FOLDERS, image_filename)
with open(image_full_path, "wb") as image_file:
image_file.write(image_data)
# Create a temporary file for processing

with [Link](delete=False, suffix='.jpg') as temp_file:

temp_file.write(image_data)
temp_file_path = temp_file.name
# Match image with stored database
match_result = process_and_match_image(temp_file_path)
print('Match Result:', match_result)
if match_result == email: # Ensure it matches the provided email
# Play login success audio
audio_path = [Link]('static/audio', 'login_success.mp3')
return jsonify({'success': True, 'audio': audio_path})
else:
return jsonify({'success': False, 'redirect': url_for('index')})
except Exception as e:
print("Error:", str(e))
return jsonify({'error': 'Image processing failed'}), 500
otp_storage = {}
@[Link]('/send_otp', methods=['POST'])
def send_otp():
data = [Link]
email = [Link]('email')

40
 print(email)
otp = str([Link](1000, 9999))
print(otp)
otp_storage[email] = otp
## server = [Link]("[Link]", 587)
## [Link]()
## [Link]("your-email@[Link]", "your-password")
## [Link]("your-email@[Link]", email, f"Subject: OTP
Verification\n\nYour OTP is: {otp}")
## [Link]()
sender_email = 'manvithar32@[Link]'
sender_password = 'eqiuggostmlivfip'
receiver_email = email
host = "[Link]"
mmail = "manvithar32@[Link]"
hmail = email
sender_name= "admin"
receiver_name='user'
msg = MIMEMultipart()
subject = "OTP-Verification"
text =f"Subject: OTP Verification\n\nYour OTP is: {otp}"
## msg = MIMEText(text, 'plain')
[Link](MIMEText(text, 'plain'))
msg['To'] = formataddr((receiver_name, hmail))
msg['From'] = formataddr((sender_name, mmail))
msg['Subject'] = 'Respected sir/mam '
server = [Link](host, 587)
[Link]()
[Link]()
[Link](mmail, sender_password)
[Link](mmail, [hmail], msg.as_string())

41
 [Link]()
send="send"
return jsonify({'success': True})
@[Link]('/verify_otp', methods=['POST'])
def verify_otp():
data = [Link]
email = [Link]('email').replace(" ", "")
otp = [Link]('otp')
if otp_storage.get(email) == otp:
del otp_storage[email]
return jsonify({'success': True})
else:
return jsonify({'success': False, 'message': 'Invalid OTP'}), 400
@[Link]("/store_order", methods=["POST"])
def store_order():
try:
print("Received request at /store_order") # Debugging log
data = request.get_json()
print("Received Data:", data) # Print received data
if data is None:
return jsonify({"error": "Invalid JSON format"}), 400
product = [Link]("product")
quantity = [Link]("quantity")
price = [Link]("price")
email = a[-1] # Ensure 'a' is correctly defined
print("Extracted Values:", product, quantity, price, email) # Debugging log
if not product or not quantity or not price:
return jsonify({"error": "Missing required fields"}), 400
# Check if the database connection is working
conn = [Link](DATABASE)
cursor = [Link]()

42
 # Debug: Check if the table exists
[Link]("SELECT name FROM sqlite_master WHERE type='table' AND
name='orders';")
table_exists = [Link]()
if not table_exists:
print("Error: Table 'orders' does not exist.")
return jsonify({"error": "Database table missing"}), 500
[Link]("INSERT INTO orders (email, product, quantity, price) VALUES (?,
?, ?, ?)",
(email, product, quantity, price))
[Link]()
[Link]()
print("Order successfully stored!")
return jsonify({"message": "Order stored successfully"})
except Exception as e:
print("Error storing order:", str(e))
return jsonify({"error": str(e)}), 500
class PDF(FPDF):
def header(self):
self.set_font("Arial", "B", 16)
[Link](200, 10, "Supermarket1 Bill", ln=True, align='C')
[Link](10)
def generate_pdf_bill(order_details, total_price, bill_id, qr_filename):
pdf = PDF()
pdf.add_page()
pdf.set_font("Arial", size=12)
# Display payment message
[Link](200, 10, "Use this QR code to pay your bill:", ln=True, align='C')
[Link](5)
# Add QR Code
[Link](qr_filename, x=75, y=30, w=60) # Adjust the size and position

43
 [Link](60) # Move down to avoid overlapping
[Link](200, 10, f"Bill ID: {bill_id}", ln=True)
[Link](5)
[Link](50, 10, "Product", 1)
[Link](30, 10, "Quantity", 1)
[Link](40, 10, "Price", 1)
[Link]()
for order in order_details:
[Link](50, 10, order[0], 1)
[Link](30, 10, str(order[1]), 1)
[Link](40, 10, f"Rs. {order[2]}", 1)
[Link]()
[Link](120, 10, "Total", 1)
[Link](40, 10, f"Rs. {total_price}", 1)
bill_pdf = f"bill_{bill_id}.pdf"
[Link](bill_pdf)
return bill_pdf
def send_email(user_email, bill_pdf):
sender_email = 'manvithar32@[Link]'
sender_password = 'eqiuggostmlivfip'
msg = EmailMessage()
msg["Subject"] = "Your Supermarket Bill"
msg["From"] = sender_email
msg["To"] = user_email
msg.set_content("Please find your bill attached. Use the QR code in the bill to make
your payment.")
with open(bill_pdf, "rb") as f:
msg.add_attachment([Link](), maintype="application", subtype="pdf",
filename=bill_pdf)
with smtplib.SMTP_SSL("[Link]", 465) as server:
[Link](sender_email, sender_password)

44
 server.send_message(msg)
@[Link]("/generate_bill")
def generate_bill():
conn = [Link](DATABASE)
cursor = [Link]()
[Link]("SELECT product, quantity, price FROM orders")
order_details = [Link]()
[Link]("SELECT SUM(price) FROM orders")
total_price = [Link]()[0] or 0
if total_price == 0:
[Link]()
return jsonify({"message": "No orders found!"})
bill_id = f"BILL{int(total_price*100)}"
# Generate QR Code
qr_data = f"Bill ID: {bill_id}\nTotal: Rs. {total_price}\nScan to Pay"
qr = [Link](qr_data)
qr_filename = f"qr_{bill_id}.png"
[Link](qr_filename)
# Generate PDF with QR Code
bill_pdf = generate_pdf_bill(order_details, total_price, bill_id, qr_filename)
# Send email
user_email = a[-1] # Replace with actual user email
send_email(user_email, bill_pdf)
# Clear order table after bill generation
[Link]("DELETE FROM orders")
[Link]()
[Link]()
return jsonify({
"message": "Bill generated and emailed successfully! Order table cleared.",
"total_price": total_price,
"bill_id": bill_id,

45
 "qr_code": qr_filename
})
@[Link]('/home1')
def home1():
return render_template('[Link]')
# Home Page Route
@[Link]('/home')
def home():
if 'email' in session:
return render_template('[Link]', email=session['email'])
return redirect(url_for('index'))
if __name__ == '__main__':
[Link](debug=False)

46
6.2.2 SCREENSHOTS

Figure.6.2.1 HOME PAGE

47
Figure.6.2.2 REGISTRATION PAGE

48
Figure.6.2.3 SUPERMARKET DASHBOARD

49
Figure.6.2.4 SUPERMARKET BILL

50
6.3 EVALUATION METRICS
The Evaluation Metrics section critically assesses the performance and efficiency
of the Intelligent Retail Experience Shopping Project. By defining and analyzing key
metrics such as checkout speed, system responsiveness, customer engagement, and
inventory accuracy, this section provides both quantitative and qualitative insights
into the system’s effectiveness. The discussion highlights areas of success, identifies
challenges, and explores potential improvements, ensuring a comprehensive evaluation
of the project's impact on the retail shopping experience.

6.3.1 QUANTITATIVE METRICS

This subsection focuses on measurable performance indicators that assess
the project’s efficiency and reliability:

• Checkout Speed

o Metric Definition: The time taken to complete the automated checkout

process without manual intervention.
o Measurement: Average checkout completion time, measured in seconds.
o Analysis: Faster checkout times enhance customer convenience, reducing
wait times and improving overall shopping efficiency.

• System Responsiveness

o Metric Definition: The ability of the system to quickly respond to customer

interactions, such as scanning items, processing recommendations, or handling
checkout transactions.
o Measurement: Latency (in milliseconds), reflecting the delay between user
actions and system responses.
o Analysis: A highly responsive system enhances the shopping experience,
ensuring smooth navigation and real-time interactions
51
● Inventory Accuracy:
o Metric Definition: The precision of the smart inventory system in tracking

stock levels and updating real-time availability.

o Measurement: Accuracy percentage calculated based on stock mismatches
vs. actual inventory levels.
o Analysis: Higher accuracy ensures better inventory management, preventing
out- of-stock issues and improving operational efficiency.

● Customer Engagement Rate:

o Metric Definition: The level of interaction between customers and smart retail

features, such as personalized recommendations and interactive displays.

o Measurement: Percentage of customers who engage with AI-driven suggestions
or use automated features.
o Analysis: Increased engagement indicates higher acceptance and effectiveness of
smart retail technologies.

● Verification Accuracy:
o Metric Definition: The precision of the system in correctly identifying

products, validating customer purchases, and ensuring accurate billing.

o Measurement: Accuracy percentage, calculated based on the number of
correctly recognized transactions vs. total transactions.
o Analysis: The system's verification accuracy is assessed based on three key
factors: Product Recognition Accuracy, ensuring correct identification of scanned
or AI- detected products; Billing Precision, verifying that customers are charged
correctly without errors; and Fraud Detection Efficiency, evaluating the system's
ability to prevent unauthorized transactions

52
 6.3.2 QUALITATIVE METRICS

This subsection explores non-measurable factors, providing insights into the

system’s impact, challenges, and future improvements:

• Areas of Success:

o Analysis: Identifying aspects of the project that have performed well, such as
seamless automated checkout, effective product recommendations, or positive
customer feedback.
o Insights: Understanding what works allows replication of successful strategies in
future enhancements.

• Challenges Faced:

o Analysis: Highlighting difficulties encountered during implementation, such as

hardware integration challenges, customer adoption resistance, or system scalability
issues.
o Insights: Recognizing these challenges helps in developing strategies for
improvement and future scalability.
• Potential Enhancements:

o Analysis: Exploring areas where the project can be enhanced, including scalability
improvements, additional features, or integration with emerging technologies.

o Insights: Identifying enhancement opportunities contributes to the project's long-

term sustainability and adaptability.

53
 CHAPTER 7
TESTING
7.1 GENERAL
The Testing phase in the Intelligent Retail System is a crucial step to ensure the
accuracy, efficiency, and security of the system. A multi-layered testing approach is
adopted to verify that the system performs optimally under various scenarios. Unit Testing
is conducted to validate individual components, such as product scanning, pricing
algorithms, and AI-based recommendation systems. Integration Testing ensures seamless
communication between different modules, such as the POS system, inventory
management, and user interfaces. User Acceptance Testing (UAT) focuses on real-world
usability by evaluating the system from a customer and retailer perspective to ensure it
meets operational requirements. This thorough testing process is vital for delivering a
reliable, error-free, and user-friendly retail solution.

7.1.1 STRATEGIC APPROACH TO SOFTWARE TESTING

The software engineering process can be viewed as a spiral. Initially system
engineering defines the role of software and leads to software requirement analysis where
the information domain, functions, behavior, performance, constraints and validation
criteria for software are established. Moving inward along the spiral, we come to design
and finally to coding. To develop computer software we spiral in along streamlines that
decrease the level of abstraction on each turn.

A strategy for software testing may also be viewed in the context of the spiral. Unit
testing begins at the vertex of the spiral and concentrates on each unit of the software as
implemented in source code. Testing progress by moving outward along the spiral to
integration testing, where the focus is on the design and the construction of the software
architecture. Talking another turn on outward on the spiral we encounter validation testing
where requirements established as part of software requirements analysis are validated
against the software that has been constructed.

54
7.2 UNIT TESTING
Unit testing focuses verification effort on the smallest unit of software design, the

module. The unit testing we have is white box oriented and some modules the
steps are conducted in parallel.

WHITE BOX TESTING

This type of testing ensures that
o All independent paths have been exercised at least once
o All logical decisions have been exercised on their true and false sides
o All loops are executed at their boundaries and within their operational bounds
o All internal data structures have been exercised to assure their validity.
To follow the concept of white box testing we have tested each form .we have
created independently to verify that Data flow is correct, All conditions are exercised to
check their validity, All loops are executed on their boundaries.

7.3 BASIC PATH TESTING

Established technique of flow graph with Cyclomatic complexity was used to
derive test cases for all the functions. The main steps in deriving test cases were:
Use the design of the code and draw correspondent flow graph.
Determine the Cyclomatic complexity of resultant flow graph, using formula: V(G)=E-
N+2 or
V(G)=P+1 or
V(G)=Number Of Regions

Where V(G) is Cyclomatic complexity, E is the number of edges,

N is the number of flow graph nodes, P is the number of predicate nodes.
Determine the basis of set of linearly independent paths.

55
7.4 INTEGRATION TESTING
Integration Testing in the Intelligent Retail System is crucial to ensuring seamless
interaction between different modules, including the user interface, backend processing,
inventory management, and AI-driven analytics. This phase validates how effectively
these components work together to deliver a smooth retail experience.

During Integration Testing, simulated retail scenarios—such as product scanning,

inventory updates, and AI-based recommendations—are executed to verify data flow
and ensure that customer actions trigger the correct system responses. Testing focuses
on data consistency, system responsiveness, and real-time processing of transactions
across different retail functions.

Additionally, real-world conditions are emulated to assess system performance under

various workloads, including peak shopping hours. This helps identify potential
bottlenecks, communication breakdowns, or inconsistencies in data flow. By ensuring
smooth collaboration between all system elements, Integration Testing strengthens the
reliability and efficiency of the Intelligent Retail System, enhancing both user and
retailer experience.

7.5 DATA FLOW TESTING

This type of testing selects the path of the program according to the location of
definition and use of variables. This kind of testing was used only when some local
variable were declared. The definition-use chain method was used in this type of testing.
These were particularly useful in nested statements.

56
7.6 USER ACCEPTANCE TESTING
User Acceptance Testing (UAT) in the Intelligent Retail System is a critical phase
where end-users, including store managers, sales staff, and customers, actively evaluate
the system’s functionality, usability, and real-world effectiveness. This phase ensures
that the system meets the expectations and operational needs of retail users.

During UAT, real-world retail scenarios—such as scanning products, updating

inventory, processing transactions, and generating AI-driven recommendations—are
tested.
Users interact with the system’s interface, assess the responsiveness of real-time
analytics, and validate the accuracy of AI-powered insights in optimizing customer
experience and store efficiency.

The primary objective of UAT is to gather feedback on system intuitiveness, ease of

navigation, and integration with daily retail workflows. This user-driven evaluation
helps identify usability concerns, potential enhancements, or areas requiring
optimization before full deployment.

The insights obtained from UAT contribute to refining the Intelligent Retail System,
ensuring it delivers an intuitive, efficient, and user-friendly experience. By addressing
feedback-driven improvements, the system is fine-tuned to enhance both customer
satisfaction and retailer operational efficiency.

57
7.7 LOOP TESTING

In this type of testing all the loops are tested to all the limits possible.
The following exercise was adopted for all loops:

• All the loops were tested at their limits, just above them and just below them.
• For nested loops test the inner most loop first and then work outwards.
• For concatenated loops the values of dependent loops were set with the help
of connected loop.
• Unstructured loops were resolved into nested loops or concatenated loops and
tested as above.
• Each unit has been separately tested by the development team itself and all the
input have been validated.
• Each loop is tested with zero iterations, one iteration, and multiple iterations to
verify boundary conditions.
• The deepest nested loop is tested first, ensuring that the iterations propagate
correctly outward.
• Ensures that exceptions raised within loops are correctly handled and do not cause
unexpected failures.
• The conditions under which the loop exits are validated to prevent infinite loops or
early termination. Loops are executed with large datasets to check for efficiency
and performance bottlenecks.
• If loops depend on external conditions or variables, those dependencies are tested
for stability. Ensures that each iteration maintains correct logic without unintended
alterations to variables.
• Loops are tested for scenarios where counters or index variables may overflow
their datatype limits.
• Validates correct use of break, continue, and return statements inside loops.
• If loops are executed in parallel or within multi-threaded environments, race
conditions and deadlocks are checked.

58
 CHAPTER 8

CONCLUSION AND FUTURE ENHANCEMENTS

8.1 CONCLUSION
The Intelligent Retail Experience project presents a revolutionary approach to
modernizing the shopping process through the seamless integration of cutting-edge
technologies such as facial recognition, voice assistance, and automated payment
systems. By fusing these innovations into a cohesive framework, the project redefines
the retail experience, offering customers an intuitive, efficient, and personalized journey
from entry to checkout.

The foundation of the system lies in advanced AI-driven personalization. Facial

recognition not only ensures quick and secure entry for returning customers but also
enables tailored product recommendations based on past interactions. The voice
assistance module, powered by Natural Language Processing (NLP), enhances customer
engagement, providing a hands-free and interactive way to explore product details,
promotions, and store navigation.

Furthermore, the integration of automated billing and payment systems streamlines

the checkout process, eliminating traditional bottlenecks and significantly reducing wait
times. This innovation enhances convenience while ensuring secure and frictionless
transactions. The project also leverages real-time analytics, offering retailers valuable
insights into customer behavior, product demand, and sales trends. These data-driven
insights empower businesses to optimize inventory management, enhance marketing
strategies, and improve overall operational efficiency.

Beyond its technological advancements, the Intelligent Retail Experience project

fosters a highly secure and privacy-conscious environment. By implementing secure
encryption protocols and stringent access controls, the system safeguards customer data,
ensuring compliance with privacy regulations and building consumer trust.

59
Ultimately, this project represents more than just a digital transformation—it ushers in
a new era of intelligent retail, where convenience, personalization, and efficiency
converge to create a future-ready shopping ecosystem. As retail continues to evolve, this
system stands at the forefront, poised to redefine customer experiences, optimize store
operations, and set new benchmarks in smart commerce.

8.2 FUTURE ENHANCEMENTS

In the context of the Intelligent Retail Experience, feature extraction is a crucial

process that enables the system to gather, analyze, and leverage key data from a variety
of sources to enhance customer experience, personalize services, and ensure security.
At the core of this process, facial recognition technology extracts unique features from
a customer face, such as the distance between facial landmarks, the shape of the nose,
and the contour of the jaw, creating a distinctive “face template”.

This template is compared with stored profiles to verify identity quickly and
securely as the customer enters the store. In addition to facial features, voice assistance
plays a pivotal role in this experience, where voice commands are processed to extract
acoustic and phonetic features. These features include tone, pitch, and speech patterns,
which are analyzed to understand and respond to customer queries, navigate aisles,
provide pricing information, and even suggest products based on previous purchases.

The system’s ability to interpret and process these voice features ensures a seamless
interaction between the customer and the store. Furthermore, automated payment
systems rely on feature extraction from product data—such as price, quantity, and
weight—along with customer profiles to ensure accurate billing. The real-time tracking
of purchases allows the system to automatically update the shopping cart and generate
an electronic bill. Feature extraction is thus a critical component in ensuring that the
system can respond intelligently and accurately to customer needs, all while providing
a smooth transition from shopping to payment.

60
 REFERENCES

[1] L. Yao, T. Li, R. Tong, K. Wang and L. Zhang, “Radar Self-Following Shopping
Cart Based on Multi-Sensor Fusion” in IEEE Access, vol. 11, pp. 77055-77072, 2023,
doi:10.1109/ACCESS.2023.3297889.
[2] S. Malik and D. Kim, “A Hybrid Scheduling Mechanism Based on Agent
Cooperation Mechanism and Fair Emergency First in Smart Factory” in IEEE Access,
vol. 8, pp. 227064-227075, 2020, doi: 10.1109/ACCESS.2020.3046097.
[3] M. Celaya-Echarri, L. Azpilicueta, V. Ramos, P. Lopez-Iturri and F. Falcone,
“Empirical and Modeling Approach for Environmental Indoor RF-EMF Assessment in
Complex High- Node Density Scenarios: Public Shopping Malls Case Study” in IEEE
Access, vol. 9, pp.46755-46775, 2021, doi: 10.1109/ACCESS.2021.3067852.
[4] M. Shahroz, M. F. Mushtaq, M. Ahmad, S. Ullah, A. Mehmood and G. S. Choi, “IoT-
Based Smart Shopping Cart Using Radio Frequency Identification” in IEEE Access, vol.
8,pp. 68426-68438, 2020, doi: 10.1109/ACCESS.2020.2986681.
[5] J. Xu et al., “Design of Smart Unstaffed Retail Shop Based on IoT and Artificial
Intelligence” in IEEE Access, vol. 8, pp. 147728-147737,2020,
doi:10.1109/ACCESS.2020.3014047.
[6] M. M. Taygur and T. F. Eibert, ‘‘A ray-tracing algorithm based on the computation
of (exact) ray paths with bidirectional ray-tracing,’’ IEEE Trans. Antennas Propag., vol.
68, no.8, pp. 6277–6286, Aug. 2020.
[7] Z . E. Khaled, W. Ajib, and H. Mcheick, ‘‘An accurate empirical path loss model
for heterogeneous fixed wireless networks below 5.8 GHz frequencies,’’ IEEE Access,
vol. 8, pp. 182755–182775, 2020.
[8] L. Azpilicueta, P. Lopez-Iturri, J. Zuñiga-Mejia, M. Celaya-Echarri, F. A.
Rodríguez- Corbo, C. Vargas-Rosales, E. Aguirre, D. G. Michelson, and F. Falcone,
‘‘Fifth-generation (5G) mmWave spatial channel characterization for
urban Environments’ system analysis,’’Sensors, vol. 20, no. 18, p. 5360, Sep.
2020.

61
[9] L. Azpilicueta, P. Lopez-Iturri, J. Zuñiga-Mejia, M. Celaya-Echarri, F. A.
Rodríguez- Corbo, C. Vargas-Rosales, E. Aguirre, D. G. Michelson, and F. Falcone,
‘‘Fifth-generation (5G) mmWave spatial channel characterization for urban
Environments’ system analysis,’’Sensors, vol. 20, no. 18, p. 5360, Sep. 2020.
[10] F. Casino, P. Lopez-Iturri, E. Aguirre, L. Azpilicueta, F. Falcone, and A. Solanas,
‘‘Enhanced wireless channel estimation through parametric optimization of hybrid ray
launching-collaborative filtering technique,’’ IEEE Access, vol. 8, pp. 83070–
83080, 2020.
[11] H. Klaina, I. P. Guembe, P. Lopez-Iturri, J. J. Astrain, L. Azpilicueta, O. Aghzout,
A. [Link], and F. Falcone, ‘‘Implementation of an interactive environment with
multilevel wireless links for distributed botanical garden in University campus,’’ IEEE
Access, vol. 8,pp. 132382–132396, 2020.

62