Image retrieval method based on characteristic enrichment area set

Abstract

The invention discloses an image retrieval method based on a characteristic enrichment area. The method comprises the following steps of firstly, acquiring a candidate characteristic point set by calculating a Hessian matrix and non-maximum value restraint, and acquiring a sub-pixel-level characteristic point set by utilizing a three-dimensional linear interpolation method; secondly, calculating a distribution matrix and an adaption matrix of characteristic points according to coordinate positions of the obtained characteristic points of an image, and by utilizing a maximum sub-matrix and an algorithm, solving a sub-matrix of the adaption matrix, namely the most dense distribution area of the characteristic points as the characteristic enrichment area of the image; thirdly, selecting a shape bottom layer characteristic, a texture bottom layer characteristic and a color bottom layer characteristic for the characteristic enrichment area; finally, measuring similarities according to a Gaussian non-linear distance function, and quickly retrieving the image according to the ascending order of the similarities. According to the method, the calculation complexity of image retrieval can be effectively reduced, and the operation efficiency and the accuracy of the image retrieval are improved.

Description

technical field [0001] The invention belongs to the technical field of image retrieval, and mainly relates to an image retrieval method based on feature-enriched regions. Background technique [0002] The research on image retrieval can be traced back to the 1970s. The early image retrieval technology is based on image text annotation, that is, Text-based Image Retrieval (TBIR). In the 1990s, large image databases gradually became the mainstream. If the traditional method is still used, it will bring a huge workload to the image retrieval work. In order to efficiently process a large number of images, content-based image retrieval technology (Content Based Image Retrieval, CBIR) is concerned by researchers. Different from the manual labeling of images by TBIR in the original system, the content-based retrieval technology automatically extracts the visual content features of each image as its index, such as color, texture, shape, etc. In this way, in addition to text annot...

AI Technical Summary

Problems solved by technology

[0003] Traditional image retrieval systems generally perform feature extraction on the entire image, regardless of image content. Usually, when people judge the similarity of images, they are not based on the similarity of the low-level visual features of the images, but on the objects described by the images. On the basis of the semantic understanding of events or events, it is precisely because of the difference between the basis for judging the similarity of images by humans and the basis for judging the similarity of similarities by computers that the visual information obtained by the computer and the semantics of the user's understanding of the image will be different. Gap between low-level and high-level retrieval needs due to information inconsistency

[0004] In addition, the importance of different regions in the image is different, and the background information other than the subject content of the image often occupies a large part of the image, which affects the characteristics of the main object in the process of feature extraction, which will not only make the image retrieval algorithm The calculation complexity is high, the operation efficiency is low, and it will lead to problems such as poor accuracy of retrieval results

Images