International Journal of Advances in Engineering and Management (IJAEM)

Volume 6, Issue 05 May 2024, pp: 770-774 [Link] ISSN: 2395-5252

Flower Classification System UsingMachine

Learning
1, 2
Ritika Gupta 1, Hashika 2, Ankita Chadha3
student, IV SEM ,M.C.A, DAV Institute of Engineering and Technology, Jalandhar ,Punjab ,India
3
Assistance Professor , M.C.A, DAV Institute of Engineering and Technology, Jalandhar ,Punjab ,India
----------------------------------------------------------------------------------------------------------------------------- ---------
Date of Submission: 15-05-2024 Date of Acceptance: 25-05-2024
----------------------------------------------------------------------------------------------------------------------------- ---------
ABSTRACT: Flowers are valued and used by photos, as well as by ambiguity among several
people all around the world for their fragrance, flower species sharing similar traits like colour and
medicinal qualities, and symbolic religious morphology. Another difficulty is in accurately and
significance. Flower categorization is a method for as precisely segmenting the image's bloom as
recognising and studying particular kinds of feasible.
flowers. We have made use of a variety of floral The purpose of this project is to use
datasets to recognise or categorize flowers with the Jupyter, an open- source dataflow and machine
help of machine learning by applying different learning library, to build an image classifying web
machine learning algorithms such as KNN application for classifying the flower image. Flask,
algorithm in addition to providing developers a simple way to
.Eighty percent of the data has been used for build web. The ultimate goal of this project is to
training purpose whereas twenty percent of that is design and optimize a convolutional neural
used for testing purpose. Helping botanists, network for use with flower classification, and
tourists, and environmental enthusiasts who wish to eventually build a simple classification web
learn about flowers is the aim of this method. application for internet around the trained network.
KEYWORDS: Machine Learning, Flower
Classification, KNN ,Random Forest, Logistic II. LITRATURE REVIEW
Regression. The study of flower recognition systems
has become increasingly important in recent years,
I. INTRODUCTION especially in areas like farming, nature
The initial step of this project is flower conservation, and environmental research. In the
classification, which involves grouping all the past, identifying flowers usually involved people
flowers according to their petal length, petal looking at them and deciding what they were. But
breadth, length, and width. There are many this process was slow, could make mistakes, and
different types of plants and animals in the world wasn't very efficient. So, researchers have been
today. The issue is that people don't realise how working on ways to use computers to do this job
useful they are for a variety of purposes in our daily automatically, using technology like computer
lives. vision and machine learning.
Categorization of Flower and plants is Then, with the rise of deep learning,
necessary for both biodiversity study and which is a type of machine learning, researchers
protection. Incorrect species identification can started using something called convolutional neural
result in inappropriate conservation efforts . networks (CNNs) to teach computers how to
Moreover, proficient taxonomists and those with recognize flowers .The research on flower
knowledge of botany and anthologies find precise detection systems has shown that using computers
flower classification to be an easy endeavour. to identify flowers can be really effective.
Correct flower identification is usually difficult and
time-consuming for beginners and non-experts. III. APPLICATIONS
Differentiating between flowers in image 1. Botanical Research: These are used to identify,
processing algorithms is still difficult today. This is classify, and study plant species. They help in
due to the fact that many flower species have the study of plant evolution, ecology and
comparable hues and forms. Automatic flower genetics.
recognition is further complicated by variations in 2. Horticulture and Agriculture: This aids in the
occlusions, perspectives, and scales in flower selection of suitable plants for cultivation in

DOI: 10.35629/5252-0605770774 |Impact Factorvalue 6.18| ISO 9001: 2008 Certified Journal Page 770
 International Journal of Advances in Engineering and Management (IJAEM)
Volume 6, Issue 05 May 2024, pp: 770-774 [Link] ISSN: 2395-5252

horticulture and agriculture, taking into

account variables such as growth habits, soil
type, and climate. Additionally, it supports
breeding initiatives that create new cultivars
with desired characteristics like increased yield
or resistance to disease.
3. Ecology and Conservation : Through the
identification of threatened species and
evaluation of ecosystem health, flower
classification aids in the monitoring and
preservation of biodiversity. It facilitates the
identification of invasive species and the
comprehension of their effects on native
plants.
4. Pharmacology and Medicine: Classification
systems facilitate the identification and study
of plants for their potential therapeutic
applications, as some flowers have medicinal
qualities. Comprehending the phytochemistry
and taxonomy of therapeutic flowers is
essential for the development of new drugs.
5. Aesthetic and Recreational Purposes: Systems
for classifying flowers are used in gardening,
landscaping, and floral design to produce
arrangements that are visually appealing. They
assist experts and hobbyists in choosing
complementary plant species for floral
arrangements,parks, and gardens.

IV. VISUALIZATION
A. Flow Chart:
FIG 1. Process of Data
All flowcharts are is a pictorial depiction
of the data or algorithm to aid in the visual
comprehension of the code. Below are the steps
B. Graph:
describing the whole process from beginning to end
A graph is a graphic or diagramtic
pictorically.
representation of data or values that is organised
and used in mathematics. This graph represents the
comparision between the sepal_length ,
petal_length , sepal_width and petal_width with
respect to iris verginica , iris setosa and iris
versicolor.

FIG 2. Comparison b/w attributes

DOI: 10.35629/5252-0605770774 |Impact Factorvalue 6.18| ISO 9001: 2008 Certified Journal Page 771
 International Journal of Advances in Engineering and Management (IJAEM)
Volume 6, Issue 05 May 2024, pp: 770-774 [Link] ISSN: 2395-5252

V. METHODOLOGY ,petal [Link] objective is to categorise iris

flowers into three species (setosa , versicolor , or
A. Datasets: virginica) based on the length and breadth of the
For our project we have taken Iris flower sepals and petals. Each of the three classes in the
data set" which is a multivariate dataset .Each of iris data collection correspond to different iris plant
the three Iris species has fifty samples in the data types which has fifty occurrences.
set. There exists four attributes of flowers which
includes sepal length ,sepal width ,petal length

ID Sepal Sepal Petal Petal Species

Length Width Length Width
1 7.1 3.5 7.9 2.1 Iris-
verginica
2 9.8 8.7 1.4 5.0 Iris-
setosa
3 0.2 3 4 5.0 Iris-
versicolor
4 5.1 8.7 4.5 6.7 Iris-
versicolor
5 0.1 0.2 11 1.9 Iris-setosa
6 11 13 16 18 Iris-
verginica
7 4.7 5.8 3.5 5.7 Iris-
versicolor
8 0.2 1.8 2.2 4 Iris-setosa
9 5.7 5.5 6 5 Iris -
Verginica
TABLE 1. Dataset of Iris Flower

B. How System Works of a number of decision trees from various

1) KNN Algorithm : K -Nearest Neighbor, is subsets of a given dataset to increase that
based on the supervised learning dataset's predictive accuracy. The random
[Link] K-NN method classifies a new forest predicts the final result based on the
data point based on similarity after storing all majority of predictions from each tree rather
the relevant data. This indicates that the K- than relying on a single decision tree
NN algorithm can quickly classify newly
discovered data into a well-suited category. 4) Logistic Regression: Logistic regression, on
the other hand is also a machine learning
2) Confusion Matrix Algorithm : Confusion algorithm. More precisely, it is a supervised
matrix is a way to show how many machine learning algorithm. It learns from
occurrences, depending on the model's labelled datasets and converts the data points
predictions, are accurate and erroneous. It into the most optimal logistic functions. This
is frequently used to assess how well can be used to predict new data sets.
categorization models which seek to
assign a category label to each C. Experimental Setups
occurrence of input perform. The iris dataset, which we gathered, has
150 tuples with the four floral properties of petal
3) Random Forest Algorithm : A Random length, petal width, sepal length, and sepal length.
Forest is a classification that uses an average The setosa, versicolor, and virginca flower species

DOI: 10.35629/5252-0605770774 |Impact Factorvalue 6.18| ISO 9001: 2008 Certified Journal Page 772
 International Journal of Advances in Engineering and Management (IJAEM)
Volume 6, Issue 05 May 2024, pp: 770-774 [Link] ISSN: 2395-5252

are the three that are featured.

About 70% of the current dataset is made up of
training data, while the remaining 30% is testing
data.
Four distinct machine learning algorithms are used
to acquire accuracy ,with a maximum accuracy of
0.96%.

The algorithms we have used is Logistic

Regression followed by Random Forest , K –
Nearest Neighbor and Confusion matrix

The text box that serves as the user interface is the

input , where the user enters sepal length, sepal
breadth, petal length ,petal breadth of the iris
species.
FIG 5. Output Screen
After entering the input values the user will get to
know about the specie of iris flower, and this will E. Result
be the output ofinput value. Experimental results demonstrate the
effectiveness of our Flower Classification System.
The trained CNN achieved more than 90%
accuracy on test data, significantly outperforming
traditional methods. In addition, the system
demonstrated robustness to lighting, background
interference, and obstacles. These results highlight
the potential of deep learning techniques to
automate flower recognition tasks.

VI. CONCLUSION
The accuracy we get for the Logistic
Regression is similar to the Confusion Matrix
algirithm. In the future, the use of flowers in
Ayurvedic medicines will be more extensive. These
flowers are also used in beauty products. New
technologies are being designed to use these
flowers in air purifiers. There are many other
benefits of using flowers in various fields, some
more to come. Flower classification and
identification is based on text processing, taking
existing flowers and their characteristics to find
their characteristics. The user can enter the
FIG 4. Accuracy of Logistic Regression attributes of the flower or just the name and see the
advantages and disadvantages of the flower as soon
D. Design as we enter the flower data. The use of flowers
The final step is to design and code the must be known to know the benefits of medicinal
frontend of the web application using Flask in VS and other the rauptic use.
Studio Code. The task for the user is to enter the
values of the sepal length ,sepal width , petal length VII. ACKNOWLEDGEMENT
and petal width which gives you the desired We take this Opportunity to express our
classification of the flower. profound sense of gratitude to all those who helped
me in the successful completion of Flower
Classification System. We also acknowledge the
efforts of Mr. Vinay Chopra HoD, Department of
Computer Applications , DAV Institute of
Engineering and Technology for their constant

DOI: 10.35629/5252-0605770774 |Impact Factorvalue 6.18| ISO 9001: 2008 Certified Journal Page 773
 International Journal of Advances in Engineering and Management (IJAEM)
Volume 6, Issue 05 May 2024, pp: 770-774 [Link] ISSN: 2395-5252

support, suggestions and modifications to improve

the quality ofthe paper.

REFERENCES
[1]. Gandhinee Rajkumar and Sameerchand
Pudaruth ,”A Mobile App for the
Identification of Flowers using Deep
Learning “ in “International Journal of
Advanced Computer Science and
Applications “ in Vol. 14 , No.5,2023
[2]. Sunil Bhutada, [Link] and [Link] ,
“Flower Recognition using Machine
Learning “ in “International Journal of
Research in Biosciences,Agriculture and
Technology” in Vol. II , May 2021 : 67-73
[3]. Md. Mizanur Rahman Nayeem ,”Flower
Identification using Machine Learning”
Available at :
[Link]
347473708
[4]. Ayesha Gurnami , Viraj Mavani , Vandit
Gajjar and Yash Khandhediya , “Flower
Categorization Using Deep Convolution
Neural Networks” Available at :
[Link]

DOI: 10.35629/5252-0605770774 |Impact Factorvalue 6.18| ISO 9001: 2008 Certified Journal Page 774