Low code rate image compression method based on down sampling and interpolation

Abstract

The invention discloses a low code rate image compression method based on down sampling and interpolation, mainly overcoming the defects of low objective PSNR value and unclear texture edges in subjective image quality in the prior low code rate image compression. The low code rate image compression method comprises the following implementation steps: (1) carrying out Laplacian pyramid decomposition on an initial image and sampling under low-pass filtration to generate a low-frequency subband signal; (2) carrying out the adaptive direction coding of the low-frequency subband signal to generatethe compressed stream information of a low code rate compressed image; (3) carrying out the adaptive direction improving decoding of the compressed stream information of the low code rate compressedimage and generating a reconstructed low-frequency subband signal; and (4) carrying out directional wave interpolation recovery on the decoded low-frequency subband signal and generating a reconstructed image. The invention has the advantages of high objective PSNR value and more clear edge and texture details of low code rate compression and can be used for recovering real-time compressed image transmission with high quality.

Description

technical field [0001] The invention belongs to the technical field of image processing, and relates to digital image compression, in particular to a low-bit-rate image compression method based on Laplacian downsampling and directional wave interpolation, which can be used for real-time communication and post-processing of low-bit-rate images High-quality image decoding recovery. Background technique [0002] Image compression and communication often require real-time and high efficiency. In recent years, multimedia applications such as image progressive transfer coding based on multi-scale representation, multiple quality services, and image reconstruction at low bit rates have been widely used, and the development prospects are promising. Due to its excellent characteristics, such as multi-scale and multi-resolution representation, efficient energy aggregation, etc., JPEG2000 based on discrete wavelet transform (DWT) has become one of the most important standard technolog...

AI Technical Summary

Problems solved by technology

This method does not reasonably aggregate relevant information during image compression encoding, and does not predict discarded high-frequency information during image restoration and decoding. Therefore, no matter from the objective PSNR value or subjective visual effect, the compression effect at low bit rates is not very good. it is good

[0006] In addition, due to the existence of a lot of important texture information and edge direction information in natural images, the traditional JPEG2000 low-bit-rate image compression method based on discrete wavelet transform (DWT) cannot effectively deal with this problem because the edge direction cannot be completely Horizontal or vertical, while discrete wavelet transform DWT can only deal with strictly horizontal or vertical texture information, so it is not as good as directional wave transform in multi-scale geometry in processing complex texture information

Images