Multi-operator image redirection method based on compressed domain content perception

Abstract

The invention discloses a multi-operator image redirection method based on compressed domain content perception. At present, most of images stored and transmitted on a network are compressed by adopting a JPEG standard. And if an image redirection algorithm based on the pixel domain is adopted, the compressed code stream of the image needs to be completely decoded to the pixel domain, and after redirection, the image is encoded. Therefore, the complexity in practical application is greatly increased, more terminal resources can be occupied, more time is consumed, and the method is difficult toapply to real-time occasions. In order to reduce the time complexity of multi-operator image redirection in practical application, the invention provides a multi-operator method for image redirectionin a DCT domain. According to the method, only an image code stream needs to be partially decoded to a DCT domain, and image redirection is carried out by using data of the DCT domain of an image; moreover, the integrity and the aspect ratio of the salient region of the image can be effectively kept, and the effect equivalent to that of a multi-operator image redirection method for keeping the aspect ratio of an important content region is achieved.

Description

Technical field [0001] The present invention belongs to the technical field of image processing, and more specifically, relates to a multi-operator image redirection method based on compression domain content perception. Background technique [0002] The current image redirection algorithms based on content perception are mostly performed in the pixel domain, such as bending algorithms, line clipping and improved algorithms, and multi-operator algorithms. Although these algorithms have a good redirection effect, in practical applications, the compressed code stream transmitted and stored needs to be completely decoded into the pixel domain and then redirected, and then fully encoded, which greatly increases the practical application. Computing time and computing resource consumption. At present, there is only a small amount of research work on image redirection based on DCT (Discrete Cosine Transform) domain. In 2009, Choi et al. proposed a method for image redirection using a ...

AI Technical Summary

Problems solved by technology

However, this algorithm only considers the so-called gradient of the image, and cannot protect the visually significant area of ​​​​the image well. At the same time, the seam in the unit of DCT block is prone to jagged effect.

To sum up, these algorithms only roughly detect the visually significant areas in the image. When there are too many deleted lines, the lines tend to pass through the important areas of the image, which significantly reduces the image quality.

Images