Channel Coding of Weighted Probabilistic Model

<p>In this paper a brand new channel error detection and correction method is provided. Artificially adding symbols to the source binary sequence makes the binary sequence present a regularity. The receiver can use this rule to implement error detection and correction. Since many redundant symbols are added, a weighted probability model lossless coding method is proposed in this paper to remove redundant information, and the reasons why Markov chains and conditional probabilities are not feasible are given. It is proven that the method in this paper can reach the channel capacity when the code length approaches infinity. It is shown experimentally that when the code rate is 1/2 in the BIAWGN channel, the method in this paper is superior to the polar code and LDPC code.</p>

Channel Coding of Weighted Probabilistic Model

<p>In this paper a brand new channel error detection and correction method is provided. Artificially adding symbols to the source binary sequence makes the binary sequence present a regularity. The receiver can use this rule to implement error detection and correction. Since many redundant symbols are added, a weighted probability model lossless coding method is proposed in this paper to remove redundant information, and the reasons why Markov chains and conditional probabilities are not feasible are given. It is proven that the method in this paper can reach the channel capacity when the code length approaches infinity. It is shown experimentally that when the code rate is 1/2 in the BIAWGN channel, the method in this paper is superior to the polar code and LDPC code.</p>

Extended Multi WLS Method for Lossless Image Coding

In this paper, the most efficient (from data compaction point of view) and current image lossless coding method is presented. Being computationally complex, the algorithm is still more time efficient than its main competitors. The presented cascaded method is based on the Weighted Least Square (WLS) technique, with many improvements introduced, e.g., its main stage is followed by a two-step NLMS predictor ended with Context-Dependent Constant Component Removing. The prediction error is coded by a highly efficient binary context arithmetic coder. The performance of the new algorithm is compared to that of other coders for a set of widely used benchmark images.