Nutrition Analyzer: AI- Powered Dietary Insights for

Enhanced Well-being
Submitted By
Guttikonda Priyanka (22H71A0516)
Nekkanti Amulya Thayi Padmavathi
(22H71A0539) Chintala Mahalakshmi
(23H75A0501) Lingamgunta Lokesh Kumar
(22H71A0524)

Under the Esteemed Guidance of

Ms. B. V. N Praveena
Assistant Professor

DEPARTMENT OF COMPUTER SCIENCE AND

ENGINEERING
Contents
• Introduction
• Problem Statement
• Literature Survey
• Research Gap
Identification
• Work Progress
• Future Work & Timeline
Introduction
• The increasing focus on health, nutrition, and preventive wellness has fueled the need for intelligent dietary
assessment tools. Nutrition science forms the foundation for understanding how food intake impacts
metabolic health, energy balance, and long-term well-being. However, tracking nutritional values and
evaluating food choices remain challenging for most individuals due to manual calculations, lack of
awareness, and time constraints.

• Recent technological advancements and AI-driven decision support systems have shifted the paradigm toward
personalized nutrition. Current trends highlight the use of AI, machine learning, and analytics to classify food,
predict health implications, and recommend healthier dietary alternatives. These innovations are transforming
how people monitor their diets and make informed nutritional decisions.
Introduction
• With rising concerns over lifestyle diseases such as obesity, diabetes, and cardiovascular disorders, real-world
applications of AI-based nutrition tools are becoming increasingly relevant. They enable users to receive
instant insights, dietary recommendations, and visual analytics to improve eating behavior. The Nutrition
Analyzer project aligns with these trends by offering a smart, user-friendly platform that evaluates nutritional
content, classifies food health levels, and suggests healthier alternatives, ultimately supporting improved
dietary habits and long-term wellness.
Problem Statement
• Existing dietary tracking methods are manual, time-consuming, and prone to inaccurate
nutritional interpretation.

• Most available systems lack instant health classification, personalized recommendations, and
healthier alternatives for informed decisions.

• There is a need for an intelligent tool that analyzes nutritional values and provides actionable wellness
insights.

• This project aims to motivate healthier eating habits by leveraging AI-based nutrition analysis
and recommendation capabilities.
Objectives
•To study existing nutrition and dietary analysis systems.
•To review current research trends in AI-driven dietary assessment.
•To understand nutritional evaluation methodologies used in practice.
•To examine the strengths and limitations of existing dietary tools.
•To analyze classification, scoring, and recommendation approaches.
•To identify gaps in personalization, feedback, and user interaction.
•To define an AI-enabled framework for nutritional analysis and health
classification.
•To develop a personalized dietary recommendation system.
•To design an intuitive platform that enhances healthy eating decisions.
Literature Survey –
Overview
•Number of Papers Reviewed: 6
•Time Period Considered: 2020 – 2023
•Sources Consulted: IEEE, Springer,
Elsevier
•Evaluation Parameters:
•Food classification approaches
•Nutrient scoring & analysis techniques
•Recommendation models
•Personalization factors
•Application usability & health relevance
Literature Survey – Paper
Template
Author(s), Year
Problem
Addresse
Methodology
/ Algorithm
Dataset /
Tools Used
Results Limitations
d
Lack of Improved user
Machine Limited
personalized diet
Learning-base Nutrition personalizatio
Li et al., nutrition recommendation
d nutrition datasets + n for disease-
2023 (IEEE suggestions s & accuracy
recommender & ML toolkit specific
Access) based on in nutrient
scoring model requirements
food intake scoring

High
Difficulty in AI-based food Custom
recognition Low
automatic classificatio food
Kaur & Gupta, accuracy performance
nutritional n nutrition
2022 and for
estimation + dataset &
(Elsevier) meaningful mixed/composit
from food nutrient Python
nutrient e meals
input estimatio ecosystem
breakdown
n model
Automated
CNN-based Faster Accuracy
dietary Image
Lee et al., food dietary depends
monitoring datasets
2021 recognition & tracking with heavily on
from visual (Food-101,
(Springer) nutrient visual image quality
food data custom)
mapping recognition
Literature Survey – Paper
Template
Author(s)
, Year
Problem
Addresse
Methodology
/ Algorithm
Dataset /
Tools
Results Limitations
d Used

Scoring & Focused only

Nutrition Clinical Generated
Zou & constraint- on diabetic
needs for nutrition diet plans
Han, 2020 based profiles,
diabetes-awar datasets, tailored for
(IEEE recommendatio lacks
e diet scoring diabetic
JBHI) n model generic
planning engine patients
usage
Enhanced
Profile-base
Lack of user- Food acceptabilit Limited
Abhinav d
profile nutrition DB y of food consideration
et al., recommender
driven + user suggestions of
2022 system with
healthy food profile based on health-risk
(Springer) preference
suggestions matrices user habits metrics
weighting
AI & Significant High
Need for AI Medical
Nakib et statistica improvement computationa
in datasets +
al., 2023 l models in prediction l overhead &
personalized nutritiona
(Elsevier) for health of dietary limited
diet & health l tables
monitoring health real- time
tracking
outcomes support
Identified Research Gaps
• From the reviewed literature, several limitations were observed in existing dietary analysis systems. Most
models primarily focus on food recognition or nutrient estimation, while lacking real-time health

classification, personalized dietary recommendations, and healthier food alternatives. Additionally, many

systems do not account for user profiles, lifestyle factors, or disease-specific nutritional needs, resulting

in limited personalization.

• Unaddressed challenges include integrating nutritional scoring with recommendation engines, providing

instant wellness insights, and enabling flexible user-controlled inputs when image data is unavailable.

Furthermore, current frameworks often exhibit computational overhead, poor interpretability, or

restricted applicability to specific user groups.

Identified Research Gaps
• These gaps highlight the scope for improvement in designing an AI-powered nutrition analysis system that
provides actionable feedback, nutritional profiles, and supportive alternatives for healthier

decision-making. Addressing these research gaps forms the motivation for the proposed Nutrition Analyzer

project, aimed at enhancing dietary awareness and promoting long-term well-being through intelligent

insights.
Dataset Used
● The system uses the Food Nutrition Dataset sourced from Kaggle, containing detailed nutritional values
of various food items.
● The dataset includes features such as calories, proteins, carbohydrates, fats, and other nutrient attributes
used for health classification.
● It is a structured tabular dataset suitable for machine learning-based food health prediction and
recommendation.
● [Link]
Working of Machine Learning Algorithms in Our Project
● In our project, multiple Machine Learning algorithms were trained to classify food items based on their

nutritional values into different health categories.

● The dataset was preprocessed using feature engineering, scaling (StandardScaler), stratified train-test

split, and SMOTE for class balancing before model training.

● The following algorithms were implemented and evaluated:

○ Logistic Regression

○ Decision Tree

○ Gradient Boosting

○ Random Forest (Selected Best Model)

Evaluation Metrics of Machine Learning Models
Model Train Accuracy Test Accuracy Precision Recall F1 Score

Random Forest * 0.9809 0.9633 0.9647 0.9633 0.9636

Gradient Boosting 0.9500 0.9224 0.9315 0.9224 0.9239

Decision Tree 0.9425 0.9174 0.9201 0.9174 0.9181

Logistic Regression 0.8681 0.8165 0.8158 0.8214 0.8158

Project Architecture
UML - use case diagram
UML - class diagram
UML - Sequence diagram
UML - Sequence diagram
UML - Activity diagram
UML - component diagram
Work Progress
• Problem identification – Completed
• Literature survey – Completed
• Gap analysis – Completed
• Methodology design – Completed
• Data Set Collection - Completed
• Implementation – Completed
Future Work
• Extend the system to support image-based food recognition using deep
learning for automatic nutrient estimation.

• Enhance personalization by incorporating medical conditions and

advanced user profiling for disease-specific recommendations.

• Integrate with wearable devices and fitness applications to provide

dynamic, activity-based diet adjustments.

• Deploy the solution as a scalable cloud-based web/mobile platform with

real-time analytics and API integration.
Conclusion
• The literature review highlights significant advancements in AI-driven nutrition analysis, including food
classification, nutrient estimation, and dietary recommendation techniques. However, current systems remain
limited in delivering real-time health classification, personalized feedback, and healthier food alternatives.
Several studies also demonstrate constraints related to dataset applicability, user profiling, and computational
[Link] findings reveal key research gaps in personalization, actionable insights, and wellness-
oriented decision support, establishing the need for a more comprehensive nutrition analysis framework.
Addressing these gaps validates the direction of the proposed project and positions it for further
[Link] a clear understanding of existing challenges and improvement opportunities, the project is
now well-prepared to proceed to the system design and implementation phase, focusing on enhanced user
experience, health metrics, and intelligent dietary recommendations.
 References
•B. Li et al., “Personalized Nutrition Recommender System Using Machine Learning,” IEEE Access, 2023.
• S. Kaur & A. Gupta, “AI-Driven Food Classification and Nutritional Estimation,” Elsevier – Expert
Systems with Applications, 2022.

•J. Lee et al., “Food Image Recognition with Nutrient Analysis for Dietary Monitoring,”
Springer–
Multimedia Tools and Applications, 2021.
• M. Zou & T. Han, “Diabetes-Aware Nutrition Recommendation Model Using Scoring Algorithms,”
IEEE Journal of Biomedical and Health Informatics, 2020.

•G. Abhinav et al., “Health-Based Food Recommendation Framework Using User Profiles,” Springer –
Health Information Science and Systems, 2022.
• F. Nakib et al., “Artificial Intelligence in Personalized Diet and Health Management,” Elsevier –
Computers in Biology and Medicine, 2023.