TY - JOUR
T1 - An enhanced dynamic time warping model for improved estimation of DTW parameters
AU - Yaniv, Ran
AU - Burshtein, David
N1 - Funding Information:
Manuscript received February 24, 2002; revised December 31, 2002. This work was supported by the KITE Consortium of the Israeli Ministry of Industry and Trade and by a fellowship from The Yitzhak and Chaya Weinstein Research Institute for Signal Processing at Tel Aviv University. The associate editor co-ordinating the review of this manuscript and approving it for publication was Dr. Jerome R. Bellegarda.
PY - 2003/5
Y1 - 2003/5
N2 - We introduce an enhanced dynamic time warping model (EDTW) which, unlike conventional dynamic time warping (DTW), considers all possible alignment paths for recognition as well as for parameter estimation. The model, for which DTW and the hidden Markov model (HMM) are special cases, is based on a well-defined quality measure. We extend the derivation of the Forward and Viterbi algorithms for HMMs, in order to obtain efficient solutions for the problems of recognition and optimal path alignment in the new proposed model. We then extend the Baum-Welch estimation algorithm for HMMs and obtain an iterative method for estimating the model parameters of the new model based on the Baum inequality. This estimation method efficiently considers all possible alignment paths between the training data and the current model. A standard segmental K-means estimation algorithm is also derived for EDTW. We compare the performance of the two training algorithms, with various path movement constraints, in two isolated letter recognition tasks. The new estimation algorithm was found to improve performance over segmental K-means in most experiments.
AB - We introduce an enhanced dynamic time warping model (EDTW) which, unlike conventional dynamic time warping (DTW), considers all possible alignment paths for recognition as well as for parameter estimation. The model, for which DTW and the hidden Markov model (HMM) are special cases, is based on a well-defined quality measure. We extend the derivation of the Forward and Viterbi algorithms for HMMs, in order to obtain efficient solutions for the problems of recognition and optimal path alignment in the new proposed model. We then extend the Baum-Welch estimation algorithm for HMMs and obtain an iterative method for estimating the model parameters of the new model based on the Baum inequality. This estimation method efficiently considers all possible alignment paths between the training data and the current model. A standard segmental K-means estimation algorithm is also derived for EDTW. We compare the performance of the two training algorithms, with various path movement constraints, in two isolated letter recognition tasks. The new estimation algorithm was found to improve performance over segmental K-means in most experiments.
KW - Baum inequality
KW - Dynamic time warping (DTW)
KW - Hidden Markov model (HMM)
KW - Speech recognition
UR - http://www.scopus.com/inward/record.url?scp=0038782382&partnerID=8YFLogxK
U2 - 10.1109/TSA.2003.811540
DO - 10.1109/TSA.2003.811540
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:0038782382
SN - 1063-6676
VL - 11
SP - 216
EP - 228
JO - IEEE Transactions on Speech and Audio Processing
JF - IEEE Transactions on Speech and Audio Processing
IS - 3
ER -