Smart Retail Checkout

Abstract

An integral part of Computer vision is Object detection. Object detection aids in pose estimation, vehicle detection, surveillance etc. Nowadays, we are using barcode to identify an object in retail stores. It consumes a lot of time as individual scanning of each item is required and results in long queues at the billing counter. The proposed system named Smart Retail Checkout, resolves this issue as it uses a camera to capture the image of multiple products at once. The input image is then given to a model which identifies the name and size of the products which are then used to generate a bill. The architecture of the proposed system is Faster R-CNN, which consists of two networks: Region Proposal Network(RPN) for region proposals and a network that uses these proposals to detect objects.