Evaluation of Lower Bounding Methods of Dynamic Time Warping (DTW)

Abstract

paper presents a brief review on the lower bounding(LB) methods applied on Dynamic Time Warping(DTW) till now. Apart from providing a survey on the methods, an attempt has been made to compare these methods in terms of constraints involved with these methods. Some Lower Bounding (LB) methods have better pruning power than others, some are better in terms of running time and also there are some which do introduce greater number of false dismissals than others. This work will help researchers in selecting a suitable lower bounding method for their application. The authors hope that this work will provide a scope of evaluating Lower bounding distances of DTW in the area of speech recognition and verification in general and will also help identify research topic and application in this area.