A commodity category identification method based on deep matching network

Abstract

The invention discloses a commodity category identification method based on a deep matching network. The method is generally divided into three processes: 1) Input the shelf image and the template image of the commodity to be recognized, and perform feature point matching on the template image and the shelf image to obtain a list of feature point pairs matched between the shelf image and the template image; 2) use Matched feature point pairs are listed, and the shelf image is aligned and cut to obtain a single product image; 3) For a single product image, use a deep matching network to classify it to determine the position and category of the product on the shelf image. The method of the invention uses a mobile phone camera to photograph shelves, overcomes the difficulties of large manpower consumption and long time consumption in the existing supermarket inventory method, and realizes commodity identification based on a repeated pattern method and a deep matching network.

Description

technical field [0001] The present invention relates to a product identification method, in particular to a product category identification method based on a deep matching network. Background technique [0002] Supermarket, as an indispensable shopping place in modern society, is more and more favored by consumers, and it also plays a vital role in the vigorous development of the retail industry in recent years. Commodities are the core of supermarkets, and good commodity management is an important means to improve the operating efficiency and comprehensive competitiveness of supermarkets, and is the foundation for the survival and development of supermarkets. [0003] As the carrier of goods, the shelf contains a lot of useful information in its state. For example, the quantity of goods reflects the real-time sales situation, which can be used to determine whether timely replenishment is needed, and provide an important reference for purchase and product structure adjustme...

AI Technical Summary

Problems solved by technology

[0005] In view of the shortcomings of the current supermarket inventory method, which consumes a lot of manpower and takes a long time, the purpose of the present invention is to provide a product category identification method based on a deep matching network, which can use a mobile phone camera to take pictures of the shelves, and identify the products in the shelf image according to the template image

Images