Isolated word recognition using weighted state probabilities (WSP), a new approach for recognition in noise

摘要

Recognition of speech in extreme noisy environments is a difficult task. A new approach is suggested to enhance the performance of recognition in very low SNRs. The weighted state probabilities (WSP) method considers the heuristic states pattern recognition based on the left to right HMM configuration and the standard probability of getting the given observation sequence. On a ten digits (Hebrew) recognition task, with SNR of 10 dB, the WSP has improved recognition results from 0% to 50%. It is suggested to apply the method, in conjunction with parallel model combination (PMC) enhancement algorithm, to very low SNR word spotting systems.