CALORIE DETECTION SYSTEM
(UML501)
Submitted to : Submitted by:
Dr. Raman Kumar Goyal Miet Pamecha
Pranav Jain
Problem Statement
This project aims to develop an automated Calorie Detection System using YOLOv5 and EasyOCR to identify
and extract calorie information from food labels. It provides a fast, accurate, and user-friendly tool, empowering
consumers with real-time nutritional insights to make informed dietary choices.
Introduction
• The Calorie Detection System is designed to extract calorie information from food labels using advanced
deep learning techniques. The purpose of this project is to simplify the process of identifying calorie values
directly from images of food packaging. This system leverages YOLOv5 for object detection and EasyOCR
for optical character recognition (OCR) to extract text data, offering a robust solution for real-time calorie
detection.
• With the increasing focus on health and nutrition, it has become essential for consumers to be aware of the
calorie content of their food. This project aims to automate this process by providing an efficient tool for
detecting and extracting calorie values, eliminating the need for manual searching or input.
Technology Stack
Frontend: HTML, CSS, JavaScript
Tools: LabelImg(annotation), Python
Backend: Flask (Python) Libraries: Pytorch, OpenCV
Machine Learning: YOLOv5 (You Only Look
Once Version 5), EasyOCR
Key Challenges
• Variability in packaging design, fonts, and text layouts.
• Handling diverse languages and symbols on food labels.
• Ensuring real-time performance and high accuracy across different food products.
• This project addresses these challenges by integrating advanced deep learning models and creating a user-
friendly interface for seamless calorie detection and extraction.
Future Enhancements
• Mobile App Integration for real-time calorie detection on smartphones.
• Support for multiple languages in OCR for better versatility.
• Enhanced UI/UX for a more interactive user experience.
• Integration with dietary apps for a complete nutrition analysis.
Workflow
1. User Upload 2. Preprocessing 3. Object Detection
6. Calorie Extraction 5. OCR on ROI 4. ROI Extraction
ROI
Yes
Found
7. Response to User 5. OCR on Full Image
?
No
Project Flow
3. Image Collection &
1. Project Planning 2. Environment Setup
Annotation
4. Model Training with
6. Backend Development 5. OCR Integration
annotations
7. Frontend Development 8. Testing & Debugging
Demo Video
Snapshots
THANK YOU !!
