Method and system for detection of representative area of automatic quasi object type image

Abstract

The invention discloses a method and a system for detection of a representative area of an automatic quasi object type image. The method comprises the following steps: obtaining feature points through analyzing features of signals in an image, automatically estimating a quasi object through grouping according to the relationship of plane coordinates of the feature points, and automatically detecting the most representative area in the whole image through further exploring information on the quasi object. The invention can be applied in image retrieval and search, image compression, image minimization, image tagging and the like for a large amount of image information on the international network.

Description

technical field [0001] The present invention relates to a method and system for automatically detecting a representative area of ​​an image based on a pseudo-object type, in particular to a method and a system for automatically detecting a representative area of ​​an image by using a method for generating a pseudo-object using feature points. Background technique [0002] Due to the popularization of digital cameras and the Internet in recent years, the number of digital images has increased exponentially. At the application level, the analysis of image content has become more and more important; most of the current research on image object detection is through The low-level information of the image, such as using the color change characteristics, detecting the edge or boundary in the image, performing image segmentation to generate blocks, and then extracting visual features from the blocks and using them to represent the blocks. However, since the image itself belongs to t...

AI Technical Summary

Problems solved by technology

However, since the image itself belongs to the 2D plane, for each block, how to determine which blocks belong to the same object in the image is a very difficult problem, and this method often produces many broken and meaningless cells block, some studies try to extract the texture features of the block when extracting the block features, and then combine blocks with similar textures and close positions to form objects, but this method still has many shortcomings, such as the same object in different parts The textures of the objects may be very different, or the same object may be very different due to the different shooting angles and light brightness when the image is taken. When applied to object detection, the effect is often not ideal.

[0003] Some other studies allow users to select a possible object area first, and then erode, diffuse, or use edge information to generate image blocks in the area, and then form objects in this area and extract features for further analysis. use, but this method requires user intervention, which is not so convenient to use

[0004] This shows that above-mentioned existing mode still has many deficiencies, is not really a good design, and urgently needs to be improved

Images