Matching given product in the shelf.
The problem statement here is to detect product images from a shelf of goods in a super market. The shelf may contain 100 product categories. So, we are asked to execute the detection or product matching task in two ways -
case 1: If aproduct image is given, we need to locate the products inside the shelfimage
case 2: If shelfimage is given to the system, the system should detect list of known products present in the shelf
The dataset contains 3153 shelf images, and 300 product images corresponding to 100 different categories containing 3 images for each of the individual category.
Dataset source - https://storage.googleapis.com/open_source_datasets/product_detection_from_packshots.tar.gz
The detailed implementation can be found here - https://github.com/wtfwolf7/Product-Matching/blob/e19de1f16399acdcafa686e815213bfc1576f795/implementation.ipynb