Grocery Detector MVP

Grocery Detector MVP is a mobile app that uses computer vision to detect and identify grocery items in real time. Built with React Native (Expo) and FastAPI + YOLOv5, it brings object detection to your phone for a seamless, user-friendly experience.

How It Works

The app lets users take or upload photos of grocery shelves. The backend uses a fine-tuned YOLOv5 model (with additional grocery classes) to detect and classify items in the image. Detected items are highlighted with bounding boxes and confidence scores, and results are sent back to the mobile app for display.

Fine-Tuning & Model Improvements

The YOLOv5 model was fine-tuned on a custom dataset to recognize a wider variety of grocery products, including beverages, dairy, snacks, and more. This improves detection accuracy and enables the app to work in real-world store environments.

Use Cases

• Inventory Management: Quickly scan shelves to count and identify products for stock-taking.
• Product Misplacement: Detect misplaced or out-of-place items on shelves to improve store organization.
• Accessibility: Help visually impaired users identify grocery items using their phone camera.
• Shopping Assistance: Instantly recognize products and get information while shopping.

Features

• Take photos or upload images of grocery items
• Real-time object detection using YOLOv5
• Display detected items with confidence scores
• Mobile-friendly interface