You are on page 1of 18

Presentation

on

“ ML and CNN powered nutrient analysis and culinary


exploration – NutriChef ”

Presented by:
Mr. Jeet Agrawal Mr. Raj Shekokar
Ms. Ritika Arya Mr. Aniket Kale

IT VII Sem

Department of Information Technology


TULSIRAMJI GAIKWAD-PATIL COLLEGE OF
ENGINEERING & TECHNOLOGY, NAGPUR

Session 2023-2024
12/31/2023 1
Aim
This project aims to empower users by providing accurate nutrient analysis,
ingredient analysis, and nutrient value calculation along with recipe recognition,
based on uploaded images and user-provided ingredient lists.
Objective
• Nutrient Analysis: The primary objective is to create a sophisticated system
capable of analyzing the nutritional content of various food items. By using ML and
CNN algorithms, the platform should be able to process images of food and extract
essential nutritional information such as calories, proteins, carbohydrates, fats,
vitamins, and minerals and most importantly health indicator.
• Culinary Exploration: The project aims to explore the culinary aspect of food
beyond traditional cooking and recipes.
• AI-Driven Insights: NutriChef seeks to provide AI-driven insights into the
nutritional profiles of foods.
• User-Friendly Interface: An important objective is to create a user-friendly and
intuitive interface that allows users to easily interact with the platform.
• Enhancing Nutritional Awareness: NutriChef aims to enhance users' nutritional
awareness by providing detailed nutritional information.
• Scalability and Accuracy: NutriChef should be designed with scalability and
accuracy in mind, allowing it to handle a wide variety of food items and deliver
precise nutrient analysis in real-time.
What’s the need?
• There are several thousand distinct dishes and fast foods in India, with countless
regional variations. This includes various types of street food, snacks, regional
specialties, and international fast food items offered by global chains.
• According to scientists, they have examined more than 20,000 Indian dishes in a
common thread.
• Nestle recently acknowledged that over 60% of food and beverage products
available in Indian markets are unhealthy.
• The Indian Times report points out that 42 per cent of the world's population
cannot afford a healthy diet – the figure is a huge 71 per cent for India.
• According to a research-based study, over 92% of people choose unhealthy food
options unknowingly but only 8% of participants preferred healthy food in their
daily life .
• As per a study by Lancet Journal every year more than 1.5 million people in India
die due to unhealthy diets.
• According to the survey Indian fast food industry is grew by 40% every year and
generates huge sales. The worth of Indian fast food market is US $27.57 billion by
2020.
What’s the need?

Fig. Indian Times research showing how many people prefers to


eat fast food in a week in India.
Data Flow Diagram
User Interacts with Interface

User uploads or clicks photo

Image recognition process / CNN

Identify Food Item in the Image using DL

Calculate Nutritional Information based on recipe

Retrieve Recipe for Identified food item from recipe database

Display Recipe and Nutritional information to user

Display Similar Images of Identified food item

User Interacts with the Displayed Data


System Flow

Display Nutritional analysis

Fig. Basic system flow functionality


System Flow

Fig. CNN Layer Structure


System Flow

Fig. Proposed CNN Structure versus VCG 19


Advantages
• Accurate Nutrient Analysis: NutriChef leverages ML and CNN technologies
to provide precise and reliable nutrient analysis.

• Promotes Healthy Eating Habits: By offering detailed nutritional insights,


NutriChef encourages users to make healthier food choices.

• Culinary Exploration and Diversity: NutriChef goes beyond traditional


recipe-based platforms by exploring the culinary aspect of food.

• Ingredient availability: Users can input their available ingredients, and the
system generates recipe list which can be cooked by those available
ingredients and reducing food waste.

• Health Monitoring Tool: NutriChef can be a useful tool for individuals


seeking to monitor their nutritional intake, especially those with specific
dietary requirements or health conditions.

12/31/2023 10
Advantages
• User-friendly interface: The intuitive and user-friendly interface makes the
system accessible to users of all skill levels, enhancing the overall user
experience.
• Variety of recipes: The extensive recipe database ensures a wide range of
recipe options.
• Continual updates: The project can be regularly updated to incorporate new
recipes, ingredients, and nutrition information, ensuring that users have access
to the latest culinary trends and nutritional data.
• Raises Awareness of Nutritional Gaps: NutriChef may reveal nutritional
gaps in users' diets, allowing them to address deficiencies and improve overall
health and well-being.
• Positive Impact on Public Health: If deployed on a larger scale, NutriChef
has the potential to contribute to public health initiatives by promoting
healthier eating habits.
12/31/2023 11
Software requirements

Front End : Html, CSS, JavaScript, React


Backend : Flask, NodeJs, ExpressJs, TensorFlow and Keras (DL), Scikit-learn
(ML)
Image Processing: OpenCV
Database: MongoDB Atlas (Cloud)
Tools : Anaconda, VS Code, PyCharm, Goggle Colabaratory
Libraries : ReactJS, Numpy, Pandas, Matplotlib, Glob etc
Hosting Platform : Fleek (Web 3.0 blockchain platform)
and many more….
Improved Landing Page

Fig. New Landing page of NutriChef


Result Page

Fig. Result page of NutriChef


Result Page

Fig. Result page of NutriChef


Future Scope

1. Displaying closely relative images or recipe based on user’s search output.

2. Personalized health driven suggestions based on previous inputs or nutrient

requirements.

3. Introducing chatbot into the platform for better user experience.

4. Providing video links with generated recipe for better understanding or learning of

recipe for users.


References

1] Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image
recognition. In: International Conference on Learning Representations (2015)
[2] Marin, J., Biswas, A., Ofli, F., Hynes, N., et al.: Recipe1M+: A dataset for learning cross-modal
embeddings for cooking recipes and food images. IEEE Trans. Pattern Anal. Mach. Intell (2019)
[3] Raboy-McGowan, D., Lu, S., Gonzalez, L.: RecipeNet: Image to recipe/nutritional information
generator (2020)
[4] Salvador, A., Drozdzal, M., Giro-i-Nieto, X., Romero, A.: Inverse cooking: recipe generation
from food images. In: IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
(2019)
[5] Lee, H. H., Shu, K., Achananuparp, P., Kokoh Prasetyo, P., et al.: RecipeGPT: Generative
pretraining based cooking recipe generation and evaluation system. In: Proceedings of the
International World Wide Web Conference (2020)
[6] Parvez, Md. R., Chakraborty, S., Ray, B., Chang, K.: Building Language Models for Text with
Named Entities. In: Proceedings of the 56th Annual Meeting of the Association for Computational
Linguistics (Long Papers), pp. 2373–2383. Melbourne, Australia (2018)
12/31/2023 18

You might also like