BURST-LIKE PACKET LOSS COMPENSATION USING A DTWfHMM HYBRID SCHEME

摘要：

Packet loss in lP and mobile networks leads to DSR systems performance degradation Several robustness approaches have been proposed to compensate the effect of packet loss. Many of these techniques rely on temporal correlation between successive voice frames to recover data and have demonstrated improvements for minimal packet losses. Other techniques have been applied in decoding stage to control over the unfavorable iniluence of missing frames. In this paper, we propose a DTW/HMM hybrid scheme to deal with long burst-like packet loss. For each word-level HMM, DTW is used to align the partially corrupted data with a representative utterance. Tben corresponding frames to the missing part are suggested for data reconstruction. Results have shown significant improvements over comparable methods for a Persian isolated-word recognition task in noisy conditions and wben up to 50％ of frames are lost.

关键词： distributed speech recognition hidden Markov models dynamic time warping packet loss frame reconstruction

作者: Ehsan Shokri Kojori Seyed Mohammad Ahadi

作者单位: Behavioral and Brain Sciences The University of Texas at Dallas,Richardson, TX, USA Electrical Engineering Amirkabir University of Technology Tehran, Iran

会议类型: 国际会议

会议名称: 2010 2nd International Conference on Signal Processing System(2010年信号处理系统国际会议 ICSPS 2010)

会议地点: 大连

会议语种:英文

页码: 714-717

在线出版日期: 2010-07-05（万方平台首次上网日期，不代表论文的发表时间）