01.The Problem
Household food waste is a significant environmental and economic issue. Many people forget what's in their fridge, leading to expired food and unnecessary purchases.
How might we use computer vision to make food inventory management effortless?
02.Technical Approach
We trained an object detection model using a custom dataset of common grocery items. The pipeline captures fridge contents via camera, classifies items, and tracks their freshness over time.
03.Key Features
- Real-time detection: Identifies 50+ food categories with 92% accuracy.
- Expiry tracking: Estimates freshness based on item type and storage duration.
- Recipe suggestions: Recommends meals from available ingredients via API integration.
04.Results & Learnings
92%
Detection accuracy on test set
50+
Food categories supported