From Dynamic Time Warping (DTW) to Hidden Markov Model (HMM) Final project report for ECE742 Stochastic Decision

Dynamic Time Warping (DTW) and Hidden Markov Model (HMM) are two well-studied non-linear sequence alignment ( or, pattern matching) algorithm. The research trend transited from DTW to HMM in approximately 1988-1990, since DTW is deterministic and lack of the power to model stochastic signals. In this report, I make a comprehensive literature study into this transition, and show that DTW and stochastic DTW, HMM are actually sharing the same idea of DP (dynamic programming). Some experiments are also performed to address this problem.