F.Y.B.

Tech Students’ Conference on Capstone Project 2 in between 2nd and 3rd July 2022,
Vishwakarma Institute of Technology, Pune, INDIA.

NutriCal
Yogita Narule, Mehvish Mukadam, Meenal Bharsakle, Mayur Gaikwad, Mayur Patil, Eshan
Mehendale

Department of Engineering, Sciences and Humanities (DESH)

Vishwakarma Institute of Technology, Pune, 411037, Maharashtra, India

Images of dishes generally made at home. We

have also seen that these models require longer
Abstract — People are very keen on measuring processing times, which we plan to improve in our
weight, healthy diets, and staying away from model. The data that has been utilized in a few
obesity, so there is an increasing demand for food models is also less and can detect only certain
calorie measurement. High-calorie intake may be food groups.
injurious and lead to various diseases.
Keywords — BMI calculator, BMR calculator, III. METHODOLOGY/EXPERIMENTAL
CNN model, food detection, web application. A. Flowchart
I. INTRODUCTION The flowchart below depicts the process we have
followed to make this project. Firstly, we collected
Good health comes from knowing the needs and
raw data and organized it to suit our needs. We
requirements of our bodies. This is the driving force used this dataset to train our CNN model to give
for our project. We have always been taught about us the required output.
the benefits of good health and given a roadmap of
how to reach it, but we fail to realize that We then created a Flask-based web application on
everyone’s body is different. Every individual which we deployed the CNN model.
requires different kinds of and amounts of food to
reach a healthy body.
The first rule of good health is to eat and sleep on
time. One must eat a balanced diet, which should
include vitamins, proteins, carbohydrates, and
minerals in appropriate quantity. Along with this,
one must also eat fresh fruits and drink lots of
water.
This project aims to fulfill the need to keep a
check on what we put inside our bodies. We have
integrated a CNN model that tells the user more
about what they are eating like calories and fats so
that users can decide what they would prefer to eat.

II. LITERATURE REVIEW

As we have mentioned earlier, the importance of
good health and a proper diet, we conducted a
thorough literature review of multiple research Fig. 1. Flowchart depicting project flow
papers. We found that CNN models that have been
developed do not recognize mixed images, e.g.,
 F.Y.B.Tech Students’ Conference on Capstone Project 2 in between 2nd and 3rd July 2022,
Vishwakarma Institute of Technology, Pune, INDIA.

B. Algorithm
The algorithm that we have utilized for this project
is a CNN model. CNN stands for Convolutional
Neural Network, which is a network architecture
for deep learning which learns directly from CNNs
CNN’s are particularly useful for finding patterns
in images to recognize objects. They can also be
quite effective for classifying non-image data such
as audio, time series, and signal data.

Fig. 4. Shows the output when the image is input

into the model.

V. LIMITATIONS
The limitations of the project are as follows:
1. The data set utilized in this project doesn’t
contain images of a wide range of fruits and
vegetables.
Fig. 2. Depicts the working of a CNN model
2. The web application is currently housed on a
IV. RESULTS AND DISCUSSIONS local host which means that it can not be
utilized by users.
We built a fully functioning CNN model that 3. The processing of dataset images by the model
detects a given input (in this case a fruit or a takes longer than necessary.
vegetable) and displays information like its
corresponding calories, fats, sodium content,
potassium content, and fiber per 100gms. We also VI. FUTURE SCOPE
created a website using Flask, that houses the BMI We would like to implement more functionalities in
and BMR calculator as well as the CNN model. this project, some of them being:
1. We would like to expand our database to
include more variety of fruits, vegetables, and
other foods as well, like cooked meals, etc.
2. We would like to increase the processing
speed of the CNN model to give instantaneous
results.
3. We would like to add more models that
suggest diet plans and other helpful dietary
habits that users can follow.
4. We would also like to deploy the web
application so that we can get feedback from
actual users and implement more functionality
Fig. 3. CNN model. after further testing.

VII. CONCLUSION
In conclusion, we picked out this specific topic
for our project so that we would be able to get an
insight into the inner workings of how a website is
constructed and how we can use programming
languages to display data in an accessible manner.
 F.Y.B.Tech Students’ Conference on Capstone Project 2 in between 2nd and 3rd July 2022,
Vishwakarma Institute of Technology, Pune, INDIA.

In addition to that this project has also introduced [3] Zhidong Shen, Adnan Shehzad, Si Chen, Hui
us to the python programming language and its Sun, Jin Liu, Machine Learning Based
many facets such as its utilization in constructing Approach on Food Recognition and Nutrition
the backend of the website and in the CNN model Estimation, Procedia Computer Science,
which are extremely useful in multiple fields of Volume 174, 2020, Pages 448-453, ISSN 1877-
study. 0509,
We have now, employing this project, working https://doi.org/10.1016/j.procs.2020.06.113.
with raw data and realized the significance of [4] REAL TIME OBJECT DETECTION USING
segregating and pre-processing data sets to be able MACHINE LEARINING AND OPENCV
Meghna Raj Saxena1 , Akarsh Pathak2 , Aditya
to derive accurate and viable conclusions by
Pratap Singh3 , Ishika Shukla4 1,2,3,4
visualization of clean data.
Department of Computer Science and
Our aim was also to help users who want to keep Engineering, IMS Engineering College,
track of their health with the information displayed Ghaziabad, Uttar Pradesh, India. International
on our website, which was constructed using Journal of Information Sciences and
HTML and CSS, and help combat dietary issues by Application (IJISA). ISSN 0974-2255, Vol.11,
the application of different technologies. No.1, 2019, (Special Issue) © International
Research Publication House.
VIII. ACKNOWLEDGMENT http://www.irphouse.com.
This project would not have been possible without [5] Study for Food Recognition System Using
the guidance of our project guide Prof. Yogita Deep Learning. Nareen O. M. Salim et
Narule ma’am. Her input and attention to detail al 2021 J. Phys.: Conf. Ser. 1963 012014
have been very helpful for this project. We are also [6] FooDD: Food Detection Dataset for Calorie
extremely grateful to Dr. S. S. Sawant FY CP 1 & Measurement Using Food Images Parisa
2 Coordinator, Assistant Head R & D, DESH, and Pouladzadeh1(). Abdulsalam Yassine1 , and
Dr. C. M. Mahajan Head, DESH, for giving us this Shervin Shirmohammadi1,2 1 Distributed and
opportunity to express our creativity and problem- Collaborative Virtual Environments Research
solving skills by conducting this project. We would Laboratory, University of Ottawa, Ottawa,
like to appreciate all the team members who have Canada
worked tirelessly on this project. Lastly, we would {ppouladzadeh,ayassine,shervin}@discover.u
like to thank our families and friends who have ottawa.ca 2 Colleges of Engineering and
supported us throughout this process. Natural Sciences, Istanbul Şehir University,
Istanbul, Turkey.
IX. REFERENCES [7] Food Recognition for Dietary Assessment
[1] D. Devi P. and D. Iyanar, "CNN based Nutrient Using Deep Convolutional Neural Networks
Extraction from Food Images," 2020 Fourth Stergios Christodoulidis1,2() , Marios
International Conference on I-SMAC (IoT in Anthimopoulos1,3, and Stavroula
Social, Mobile, Analytics and Cloud) (I- Mougiakakou1,4 1 ARTORG Center for
SMAC), 2020, pp. 398-402, doi: 10.1109/I- Biomedical Engineering Research, University
SMAC49090.2020.9243324. of Bern, Bern, Switzerland
[2] Janmenjoy Nayak, Kanithi Vakula, Paidi {stergios.christodoulidis,marios.anthimopoulo
Dinesh, Bighnaraj Naik, Danilo Pelusi, s, stavroula.mougiakakou}@artorg.unibe.ch 2
Intelligent food processing: Journey from Graduate School of Cellular and Biomedical
artificial neural network to deep learning, Sciences, University of Bern, Bern,
Computer Science Review, Volume Switzerland 3 Department of Emergency
38,2020,100297, ISSN 1574- Medicine, Bern University Hospital, Bern,
0137,https://doi.org/10.1016/j.cosrev.2020.100 Switzerland 4 Department of Endocrinology,
297. Diabetes and Clinical Nutrition, Bern
University Hospital, Bern, Switzerland.