ECG Compression Transmission Rate Distorted Optimal Embedded And Research Unequal Protection

In the current number of error-correcting codes, low density parity check code is one of kind of the codes which is the closest to Shannon limit until now. It has been widely used in the actual communication system, especially in information protection systems and communications systems because of its own superior performance and good correction decoding complexity. From the applications division of Low Density Parity Check(LDPC) codes, the paper focuses on the application of LDPC codes in the field of communications. Through studying and researching on the degree distribution characteristics and the decoding algorithms of LDPC codes, LDPC code is applied into unequal error protection of Electrocardiograph (ECG) signals during their transmission, to realize the application of the above systems. The main research topics and contribution are as follows:1) Achieving ECG signal coding by combining Rate-Distortion Optimized-Embedding(RDOE) with adaptive arithmetic coding algorithm.The traditional coding method is to encode the coefficients one by one until completing the entire process after the wavelet decomposition. However, when the source bit stream is truncated at a certain point, the amount of information carried by next bits will be lost, resulting in serious distortion of the reconstructed signal. Embedded coding can solve this problem very well, it takes the advantage of bit plane coding, so even the decoding bit stream is truncated, we can still recover some information to achieve reconstruction signal. The rate-distortion optimal embedded will divide source bit stream into significant identification and refinement coding, and it will be encoded as different forms only when the rate distortion slope of current bit is larger than the slope threshold of current plane. Adaptive arithmetic coding can do online and real time probability distribution statistics to bit stream.2) This paper not only uses LU decomposition method to construct different rate of regular LDPC codes, but also combines density evolution theory with methods of Gaussian approximation to constructed irregular LDPC codes to achieve unequal error protection of ECG signals in the transfer process. For irregular LDPC codes, asymmetrical nodes degree distribution of variables is the basis of the unequal error protection. Since during the iterative decoding process, the nodes with high degree will receive a relatively more parity check information from the adjacent nodes. Therefore, the final judgment will have a greater probability to decoder correctExperimental data from MIT-BIT date center, and experimental results show that:under taking account of the surrounding correlation of current symbol, the embedding algorithm based on the rate distortion has a satisfactory performance that it can gain well compression. Meanwhile, the results also show that:the use of LDPC codes can provide good protection for the important bits of ECG signals.