Adaptive Compressed Sensing Image Coding Method Based on Measurement Domain Saliency Detection Model
An image coding and compressed sensing technology, applied in the field of image coding, can solve the problems of rate-distortion performance increasing coding complexity and unfavorable wireless sensor network applications, etc., to achieve the effect of improving reconstruction quality and high rate distortion performance
Active Publication Date: 2021-12-07
XINYANG NORMAL UNIVERSITY
View PDF4 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, the above-mentioned adaptive measurement methods all use full-sampled images to extract features, and the amount of calculation introduced may be equivalent to that of full transformation. The rate-distortion performance is improved at the cost of increased coding complexity, which is not conducive to the application in wireless sensor networks.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0056] The technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.
[0057] The block diagram of the low-complexity compressed sensing image encoding and decoding proposed by the present invention is as follows: figure 1 shown. Set the total image measurement rate S, and determine the total measurement times M as
[0058] M=N×S (1)
[0059] On the encoding side, first, the scene is compressed by the imaging device with size I r × I c (N=I r · I c ) in the image x of each block implements pre-measurement, that is, the preset initial measurement times M 0 as follows:
[0060]
[0061] In the formula, round[·] is the rounding operator. Divide the image x into n blocks of size B×B, where B is 16, and the i-th image block is recorded as a column vector form x i (i=1, 2,..., n, n=N / B 2 ), the generated size is M 0 ×B 2 The Gaussian random measurement matrix Ф B0 , the calculated lengt...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses an adaptive compressed sensing image coding method based on a measurement domain saliency detection model, which includes the steps of: (1) dividing an input image x into several non-overlapping image blocks x i ;(2) Set the initial measurement times M 0 , construct the initial block measurement matrix, and perform pre-measurement on each block, and obtain the initial measurement vector y of each block 0i ; (3) Use the initial measurement vector y of each block 0i Implement saliency detection in the measurement domain, and calculate the normalized block saliency w of each block i ; (4) According to the normalized block saliency degree w i , adaptively set the number of measurements M of each block i , and according to the measurement times of each block, construct the corresponding Gaussian random measurement matrix Φ Bi , calculate the measurement vector y of each block i ; (5) The decoder receives the measurement vector y of each block i , calculate y i the length of M i , re-estimate the normalized block saliency (6) weight the objective function of the image reconstruction model with the normalized block saliency estimate, establish an adaptive global reconstruction model, and use the gradient projection method to solve the adaptive global reconstruction model, generating The final reconstructed image The present invention can effectively improve the subjective and objective reconstruction quality of the image, and its overall performance has achieved a greater rate-distortion performance improvement compared with the prior art.
Description
technical field [0001] The invention belongs to the technical field of image coding, and relates to a low-complexity coding method based on compressed sensing. In particular, it proposes realizing saliency detection in a measurement domain, and implementing compressed sensing measurement adaptively to saliency information, so as to improve the rate-distortion performance of image coding. Background technique [0002] In wireless sensor networks, the energy consumption and bandwidth of image sensor nodes are greatly limited. However, traditional image coding techniques (such as JPEG) are based on the transform coding framework, which needs to introduce a large amount of calculation to implement a full transformation of the image. Therefore, the traditional Image encoding will greatly reduce the life cycle of image sensing nodes. In order to prolong the life cycle of image sensing nodes in wireless sensor networks, a new low-complexity image coding technique needs to be found....
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06T9/00
CPCG06T9/00H04N19/147
Inventor 李然刘宏兵刘正辉
Owner XINYANG NORMAL UNIVERSITY