CS-based image optimized decoding method

Abstract

To solve the problem that the decoding quality of a conventional mainstream JPEG image coding standard is decided by transformation and quantification in an encoder, the invention brings forward a CS-based decoding method according to the CS theory. DCT of a coding end is regarded as observation of an original image signal, image sparse is carried out through total variation (TV), quantized noise is considered as observation noise, image decoding is transformed CS optimized reconstruction, and thus CS accurate reconstruction replaces inverse DCT in JPEG decoding. On the above basis, CS optimized decoding based on block merging is provided to further improve the decoding quality. Sufficient experiment results show that the scheme of the invention enables the objective and subjective quality of image decoding to be substantially improved compared with conventional JPEG decoding. Compared with the prior art, the method has the advantages that the objective and subjective quality of image reconstruction at a decoding end is greatly improved while a conventional JPEG coding end structure and a code stream format are not changed, thereby having a great significance in the development of image printing industries.

Description

technical field [0001] The invention belongs to the technical field of image printing, and in particular relates to a CS-based image optimization decoding method. Background technique [0002] In the image printing industry, there is a common problem: huge image printing requires higher and more advanced equipment, but advanced equipment alone is not enough, because a large amount of printing information is easy to be lost during the printing process, which seriously Affects printed image quality. At present, the mass digital images generated by the increasing multimedia applications have brought enormous pressure on storage and transmission. Therefore, the research on efficient digital image compression coding technology is of great significance. After a series of research, JPEG based on Discrete Cosine Transform (DCT) and JPEG2000 image compression standard based on Discrete Wavelet Transform (DWT) were finally produced. At present, the JPEG coding standard has been wide...

AI Technical Summary

Problems solved by technology

[0002] In the image printing industry, there is a common problem: huge image printing requires higher and more advanced equipment, but advanced equipment alone is not enough, because a large amount of printing information is easy to be lost during the printing process, which seriously Affects printed image quality

At present, the JPEG coding standard has been widely used, but under the traditional transform coding framework, only the inverse transform decoding corresponding to the transform at the encoding end can be used to reconstruct the image signal, resulting in a fixed and unique decoding method.

For example: in JPEG, the decoding end can only use the inverse DCT transformation corresponding to the encoding end; the diversity of image signals requires the optimal sparse space to be variable, and under the transform coding framework of JPEG, the image can only be transformed by DCT. A transforming base representation that cannot adequately sparsely represent images, eventually resulting in reduced image decoding performance

Images