Performance of Lossy P-LDPC Codes over GF(2)

This paper presents a low complexity lossy compression algorithm with protograph LDPC codes (Lossy P-LDPC Codes) over GF(2), for binary symmetric sources. The proposed scheme takes a shuffled reinforced belief propagation (SRBP) algorithm for encoding, which achieves asymptotically the theoretical rate-distortion function. Additionally, the proposed coding algorithm converges fast and performs very close to the counterparts using ultra sparse LDPC codes over GF(256). Simulation results also reveal that the code structures need to be further optimized, in order to adapt to different source statistics.