Computer Vision for Shelf Product Detection

This repository contains the implementation of a computer vision-based system to identify products on store shelves from a reference image. This system can assist visually impaired customers and automate store management tasks, such as detecting low-stock or misplaced products.

Project Overview

Objective

Develop a computer vision system that:

1. Identifies products on a store shelf from a given reference image for each product.
2. Reports:
   • The number of instances of each product.
   • The dimensions (width and height in pixels) of each instance.
   • The position (center of the bounding box) of each instance in the image reference system.

Example Output

For a store shelf image:

Product 0 - 2 instances found:
  Instance 1 {position: (256, 328), width: 57px, height: 80px}
  Instance 2 {position: (311, 328), width: 57px, height: 80px}
Product 1 - 1 instance found:
  Instance 1 {position: (400, 280), width: 50px, height: 75px}
...

Tracks

1. Track A - Single Instance Detection:

   • Develop an object detection system to identify single instances of products from a reference image and a shelf image.
   • The system should correctly identify all products displayed in the shelf image.

2. Track B - Multiple Instances Detection:

   • Extend the system from Track A to detect multiple instances of the same product.

Implementation

Inputs

• Reference Images: Images of individual products.
• Scene Image: A picture of a store shelf containing multiple products.

Outputs

• Number of instances of each product.
• Bounding box dimensions (width, height) for each instance.
• Position of the bounding box center for each instance in the image.