Novel Fault-Tolerant Decompression Method of Corrupted Huffman Files

Repair and Restoration of Corrupted LZSS Files

“…In this section, the error repair capability of LZSRD is compared with that of the method proposed in [8]. The method proposed in [8] utilizes compression coding and grammar rules to determine the error position and correct the errors. The present work uses the residual redundancy left by the LZSS encoder to carry the verification information to find and repair errors.…”

“…Every time the quantity of the injected errors (represented by bit error rate (BER)) changes, 100 experiments are performed on each compressed file. In the experiments, LZSRD and the method proposed in [8] are used to repair the errors and decompress the files. The average probability of successful error correction and decompression of the files is calculated and recorded.…”

“…7. The blue and red curves represent the average probability of the successful decompression of LZSRD and the method proposed in [8], respectively. As shown in Fig.…”

“…7, when BER = 6 × 10 −4 , the method proposed in [8] can hardly decompress the file correctly, and the probability that LZSRD can successfully decompress is approximately 0.9. Therefore, the error repair capability of LZSRD is considerably higher than that of the method proposed in [8] because the latter cannot find a replacing error that does not destroy the dictionary structure; moreover, the error bit may be corrected only when the interval between the error bits is greater than the error detection delay. As BER increases, the error bits cannot be corrected once the interval is less than the error detection delay.…”

“…Kwon et al [7] introduced three special bit patterns in compressed data and proposed an error detection method, which required no additional bits to detect the error codes, but he failed to develop a feasible error-correcting method for the corrupted compressed file. Wang et al [8] established a mathematical model for error bit detection by using compression coding rules and grammar rules, estimated the rough range of error bits, and used heuristic method to determine the exact positions of the errors and correct the errors. But the heuristic method requires grammar rules as the prior information and cannot correct the replacement errors.…”