Showing papers by "Arnon D. Cohen published in 1997"

Analysis by synthesis ECG signal compression

Abstract: The authors introduce a new ECG compression algorithm, and a new distortion measure. The distortion measure, called the Weighted Diagnostic Distortion (WDD), is based on comparing PQRST complex features (such as: RR interval, QT interval, ST elevation, etc.) of the original ECG signal and the reconstructed one. The compression algorithm is based on analysis by synthesis coding. It consists of a beat codebook, long and short term predictors, and an adaptive residual quantizer. The compression algorithm uses the WDD measure in order to encode, by means of analysis by synthesis, every beat of the original ECG signal. The compression algorithm has been applied to the MIT-BIH Arrhythmia Database. A rate of approximately 100 bits per second has been achieved with a very good quality (WDD below 4%, and PRD below 8%).

Unsupervised speaker classification using self-organizing maps (SOM)

Abstract: An algorithm for unsupervised speaker classification using Kohonen SOM is presented. The system employs 6/spl times/10 SOM networks for each speaker and for non-speech segments. The algorithm was evaluated using high quality as well as telephone quality conversations between two speakers. Correct classification of more than 90% was demonstrated. High quality conversation between three speakers yielded 80% correct classification. The high quality speech required the use of 12/sup th/ order cepstral coefficients vector. In telephone quality speech, an additional 12 features of the difference of the cepstrum were required.

Robust isolated word recognition using WSP-PMC combination.

Abstract: A new robust algorithm for isolated word recognition in low SNR environments is suggested. The algorithm, called WSP, is described here for left to right models with no skips. It is shown that the algorithm outperforms the conventional HMM in the SNR range of 5 to 20db, and the PMC algorithm in the range 0 to-9db.