Study Of Binary Arithmetic Coding And The Joint Deconing Based On VLC

With the continual development of communication technology, people put forward a higher requirement on communication quality. How to make data compression and transmission reliably and efficiently has always been a research focus. Source coding can make data transmission effective, while channel coding can make data transmission reliably. Efficient binary arithmetic coding technology can provide protection for the information transmission. Using the joint source-channel decoding (JSCD) is able to ensure the validity and improve the noise immunity, to achieve the reliability requirements of communication system.After further study of traditional binary arithmetic coding algorithm, a novel arithmetic coding algorithm based on a binary look-up table is proposed. According to the characteristics of the residual data from the predicted image coded, we combines with Exp-Golomb code table to establish a new code table of binary residual data. This code table can solve the binary redundancy problem of larger residual coefficient. To enable the binary bit-stream efficient for probability estimation, we design a binary bit-stream post-processing method to improve the probability estimation processing speed. Experimental results show that the bit rate reaches 4.388bits/pixel. Compared to CMBP algorithm, it decreases by 0.234bits/pixel.To improve the data transmission performance, we also propose an iterative JSCD algorithm based on VLC-LDPC. According to the diffuse error rate problem, firstly, we decode VLC based on the maximum a posteriori probability. Then, optimize VLC-Viterbi grid searching method based on bit to forming a new VLC-Viterbi decoding algorithm. To solve the poor noise immunity performance of VLC-Viterbi decoding, we add LDPC to channel decoding. To further improve the overall decoding performance of simple cascade connection methods, we combine the VLC-Viterbi and LDPC to iterative decode. It forms a VLC-LDPC iterative JSCD algorithm. The experimental results show that compared to separate decodes algorithms, the proposed algorithm increases by about 13 dB in PSNR when the SNR is 4dB.Arithmetic coding algorithm based on a binary look-up table can improve the compression efficiency of the residual image data.VLC-LDPC iterative JSCD algorithm can provide robust and reliable image quality in noise channel.