Quick matching method and system for similar images

Abstract

The invention relates to the field of computer image retrieval, and provides a quick matching method and a quick matching system for similar images. The method comprises the steps of background preprocessing and foreground matching. The background preprocessing step specifically comprises the sub-steps of S101, extracting a visual feature of each image to be retrieved in an image database; S102, mapping M text feature labels on each image to be retrieved in the image database; and S103, building a joint sorting index table F(i)i( / i) by using a text label as an index unit, and forming image sorting. The step of foreground matching specifically comprises the sub-steps of S201, receiving and extracting the visual feature of a query image O; S202, determining the text feature label of the query image; S203, using images in the joint sorting index table corresponding to the text feature of the query image as the alternative matching image library of the query image O; and outputting a matching result according to the matching quantity. According to the method and system provided by the invention, the search range of the image matching is greatly reduced, and the image retrieving speed is improved.

Description

technical field [0001] The invention relates to the field of computer image retrieval, in particular to a fast matching method and system for similar images. Background technique [0002] The image matching system can match the image similar to the queried image content information in the image database according to the similarity degree of the information contained in the image content, and realize the output of the image matching result. The image matching system can be used to find homologous pictures or target objects. With the rapid development of the Internet and image processing technology, the practical application value of the image matching system is increasing day by day. [0003] Image matching features mainly include text features and visual features. Among them, text features need to be manually annotated in advance. Using text features to achieve image matching can reduce the difficulty of judging image similarity. However, for image matching systems with a l...

AI Technical Summary

Problems solved by technology

Among them, text features need to be manually marked in advance. Using text features to achieve image matching can reduce the difficulty of judging image similarity. However, for image matching systems with a large amount of data, the preprocessing process of manual marking is a difficult task.

Visual features are basic features with intuitive meanings such as color, texture, and shape of an image. They correspond to the information obtained by human visual perception of lightness, color, and shape. Visual features can intuitively express the content information of an image. , but compared with text features, the similarity calculation process of visual features takes longer

Images