Multi-feature content-based image retrieval method and system

Abstract

The invention provides a multi-feature content-based image retrieval method and system. The multi-feature content-based image retrieval method comprises the steps that according to various single low-layer features of images which are obtained through extraction and serve as retrieval terms, image retrieval is executed on the basis of normalization similarity measurements according to the various single low-layer features so that corresponding single low-layer features of various stored images can be matched with the obtained single low-layer features respectively; image results obtained through retrieval are ordered in a unified mode and output on the basis of total similarity measurements of the image results obtained through retrieval. In this way, the effect that in the content-based image retrieval process, combination of multiple low-layer features is optimized in the linear weighting mode of unifying the similarity measurements of all the low-layer features is achieved, the recall factor and the pertinency factor are increased, and multi-aspect actual requirements of users for content-based image retrieval are better met.

Description

technical field [0001] This application relates to the field of retrieval, in particular to a multi-feature content-based image retrieval method and system. Background technique [0002] With the development of Internet search engines and retrieval technologies, content-based image retrieval (Content-based image retrieval, CBIR) was proposed and gradually matured. Common search engines such as Google and Baidu have also successively launched corresponding popular commercial search technologies, and CBIR has gradually entered people's daily online life. The information contained in an image often goes far beyond the text. During the search process, how to process images, analyze images, extract image features, and especially how to match two images is very important to CBIR. [0003] The existing CBIR technology is mainly based on knowledge in the fields of image processing, pattern recognition, computer vision, and image understanding, and introduces new technologies from t...

AI Technical Summary

Problems solved by technology

[0005] The main purpose of this application is to provide a multi-feature content-based image retrieval method and system to solve the problem of optimizing the combination of multiple low-level features by unifying the linear weighting method of the similarity measure of each low-level feature in content-based image retrieval , so that images similar in different features can be retrieved at the same time during retrieval, thereby improving the detection rate and precision rate; and then solving the complementary and inhibiting characteristics of the characteristics of the low-level features, and further effectively improving the content-based Various aspects of image retrieval meet the actual needs of users for image retrieval and improve user experience in content-based image retrieval

Images