Information hiding based vector quantization compression method

Abstract

The invention provides an information hiding based vector quantization compression method which can greatly improve the compression performance of a conventional vector quantization compression method. The PSNR can be up to 40 dB when the compression ratio is 16 times, so the PSNR performance of the information hiding based vector quantization compression is increased by more than 12 dB when compared with that of the conventional vector quantization compression method, the PSNR performance is increased by more than 5 dB when compared with that of a JPEG2000 compression method, and the PSNR performance can be increased by 13 dB for certain images; and the information hiding based vector quantization compression method is mainly used for transmission and storage. The information hiding based vector quantization compression method is characterized in that the information hiding technology is utilized to creatively select an image region of interest for embedding of special error correction information rather than embedding secret information in the vector quantization for joint encoding and hidden transmission, a receiving end extracts correction information from normal vector quantization index data without distortion, and the quality of a restored image is greatly improved. The information hiding based vector quantization compression method has a high-performance data compression effect and relatively simple hardware implementation, and not only can compress image data, but also can compress non-image data.

Description

technical field [0001] The invention relates to a data communication method, in particular to a vector quantization compression method based on information hiding, and belongs to the field of communication (such as data communication technology, etc.). Background technique [0002] Information hiding, also known as data hiding, is an important branch of information security. It uses human visual redundancy to embed secret information into the carrier, so as to achieve the purpose of secure transmission of secret information. [0003] Information hiding is divided into lossless information hiding or lossy information hiding. Lossy information hiding such as LSB and DCT information hiding methods, after information hiding, the carrier data has a certain loss, and the information has no loss or a certain loss. Lossless information hiding methods, such as the histogram-based information hiding method, the difference expansion information hiding method, etc., have no loss of carr...

AI Technical Summary

Problems solved by technology

In terms of objective evaluation criteria, the peak signal-to-noise ratio (PSNR) may vary a lot

[0009] (2) The PSNR of vector quantization VQ image compression needs to be improved

[0010] Due to the limitations of the learning algorithm and the storage size of the codebook, the peak signal-to-noise ratio (PSNR) of vector quantization compression is not high. For example, for 16 times image compression, the PSNR may only be about 30dB, which needs to be improved

Images