Method for simplifying similarity determination of DCT pixel grayscale pictures

Abstract

The invention discloses a method for simplifying similarity determination of DCT pixel grayscale pictures. The pixel of a picture is reduced to be a pixel of 32*32; after DCT transformation, a 8*8 calculation process is reduced; a DCT value of each pixel is obtained; and mean values of the DCT values are compared to obtain a hash. According to the invention, the 8*8 calculation process is reducedby DCT transformation, so that the pure color information is filtered and the detail expression is improved further; the similarity determination rate is about 40%; and the computing load is reduced although the similarity determination rate is affected, so that some time costs are lowered.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a method for judging the similarity of simplified DCT pixel grayscale pictures. Background technique [0002] As intellectual property rights are valued by the society, there are more and more intellectual property disputes. As an important part of intellectual property rights, pictures are frequently infringed. At present, the whole web crawler technology can obtain massive picture information, but how to select pictures similar to the picture right holder from the massive picture information is a problem that needs to be solved. With the current technology, it is difficult to use computer methods to directly judge the infringement of pictures, but it can screen a large number of pictures and exclude dissimilar pictures, saving a lot of labor time and cost of screening pictures for manual judgment. [0003] In order to solve the above problems, the applicant applied for "a metho...

AI Technical Summary

Problems solved by technology

[0003] In order to solve the above problems, the applicant applied for "a method for judging similarity of pictures" with the application number "2017111545551". This method reduces the amount of processing of color information by comparing the grayscale of two pictures; The image is divided into matrix and grayscale at the same level, which can satisfy the same photo, adjust the color proportionally, and judge the similarity of the photo after operations such as proportional scaling. However, if the grayscale average is compared, the displayed details are relatively small, and the judgment The success rate of the similar degree is low, about 33%, and finally needs to be manually judged, and the workload is slightly larger

[0004] In order to overcome the above problems, the applicant performed DCT processing on the details in another patent, adding detailed expressions, but invisibly increasing the amount of calculation

Images