NutriLens
NutriLens is a smart web app that analyzes packaged food labels to rate product healthiness on a scale of 1 to 10. It uses OCR and NLP to extract and assess ingredients and nutritional values.
Created on 5th April 2025
•
NutriLens
NutriLens is a smart web app that analyzes packaged food labels to rate product healthiness on a scale of 1 to 10. It uses OCR and NLP to extract and assess ingredients and nutritional values.
The problem NutriLens solves
NutriLens is a web-based tool that simplifies and automates the analysis of packaged food labels using OCR and NLP. It extracts text from label images and applies rule-based logic, combined with guidelines from sources like WHO and FDA, to evaluate ingredients and nutritional content. The system then generates a health score from 1 to 10, highlighting harmful additives and offering actionable insights. Designed with scalability in mind, the modular backend supports extensions like barcode scanning, user-specific dietary filtering, and real-time recommendations, making food evaluation faster, safer, and smarter for end-users. By automating the analysis and interpretation of complex food labels, NutriLens reduces manual effort, enhances food transparency, and empowers users—especially those with dietary restrictions—to make quicker, safer, and more informed choices. Increased the accuracy of the model by 90%.
Challenges we ran into
One of the biggest hurdles we faced was dealing with inaccurate or noisy OCR results, especially when food labels had curved surfaces, low lighting, or stylized fonts. This led to misread words, broken ingredient names, and missing nutritional values, which impacted the accuracy of the health scoring. To overcome this, we implemented a preprocessing pipeline using OpenCV that included: Grayscale conversion, Adaptive thresholding ,Perspective correction for skewed labels, Noise reduction filters.
Tracks Applied (1)
