A method of commodity image feature description based on multi-scale visual word bag model

Abstract

A commodity image feature description method based on a multi-scale visual word bag model solves the problem of poor commodity image classification using existing visual word bag feature description methods, and belongs to the field of image classification. The invention includes: performing wavelet multi-scale decomposition on the 2 times upsampled commodity image, extracting multi-scale feature points, performing edge filling, delineating the local area of ​​the image on the edge-filling image, using SURF local features and quantization normalization The color vector corner features of the delineated local area are described, and the final descriptor of the local area feature is obtained through serial fusion; the local area features are clustered to form word bags; each local area feature and word bag The word feature vector in is used for distance measurement, encoding the local area features, using the diagonal concentric moment space model to perform feature aggregation to form an image descriptor, and performing dimensionality reduction to obtain the visual word bag feature vector; the obtained visual word bag feature vector Classify with a classifier.

Description

Technical field [0001] The present invention relates to a multi-scale image feature merchandise package visual word description of the model and belongs to the field of image classification. Background technique [0002] Scale of my country's e-commerce continues to grow, the industry has become an important influence people's lives. As the Internet continues to progress to promote the rapid development of Taobao, Jingdong and only product other major e-commerce platform, online shopping has become the main form of public purchasing of goods, more and more recognized by people. According to China Internet Network Information Center, as of March 2020, the scale of my country was 904 million Internet users, Internet penetration rate of 64.5%, my country's online shopping users reached 710 million, 2019 transactions reached 10.63 trillion yuan, an increase of 16.5 %. To provide users with more excellent shopping experience to quickly browse items, search for the positioning, a good ...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problem that the commodity image classification effect of the existing visual word bag feature description method is not good, the present invention provides a product image feature description method based on a multi-scale visual word bag model

Images