Dynamic time warping

About: Dynamic time warping is a research topic. Over the lifetime, 6013 publications have been published within this topic receiving 133130 citations.

DTW-Based Gait Recognition from Recovered 3-D Joint Angles and Inter-ankle Distance

Abstract: We present a view independent approach for 3D human gait recognition. The identification of the person is done on the basis of motion estimated by our marker-less 3D motion tracking algorithm. We show tracking performance using ground-truth data acquired by Vicon motion capture system. The identification is achieved by dynamic time warping using both joint angles and inter-joint distances. We show how to calculate approximate Euclidean distance metric between two sets of Euler angles. We compare the correctly classified ratio obtained by DTW built on unit quaternion distance metric and such an Euler angle distance metric. We then show that combining the rotation distances with inter-ankle distances and other person attributes like height leads to considerably better correctly classified ratio.

Continuous speech recognition system

Abstract: Recognition of continuous speech by comparison with prestored isolated words may be confused by the merging together of spoken adjacent words (coarticulation). Improved recognition is attained by generating overlap-words, e.g., words whose first phoneme is the end phoneme of the preceding word in a string of words. The reference candidate series of overlap-words is transformed under dynamic time warping so as to time-match the utterance series of overlap-words.

Fast speaker adaptation for speech recognition systems

Abstract: Different speaker adaptation methods for speech recognition systems adapting automatically to new and unknown speakers in a short training phase are discussed. The adaptation techniques aim at transformations of feature vectors, optimized with respect to some constraints. Two different adaptation strategies are discussed. The first one is based on least mean-squared-error optimization. The second method is a codebook-driven feature transformation. Both adaptation techniques are incorporated into two different recognition systems: dynamic time warping (DTW) and hidden Markov modeling (HMM). The results show that in both systems speaker-adaptive error rates are close to speaker-dependent error rates. In the best case the mean error rate of four test speakers decreases by a factor of six compared to the interspeaker error rate without adaptation. A hardware realization of the speaker-adaptive HMM-recognizer is described. >

Applying Boosting to Similarity Literals for Time Series Classification

Abstract: A supervised classification method for temporal series, even multivariate, is presented. It is based on boosting very simple classifiers, which consists only of one literal. The proposed predicates are based in similarity functions (i.e., euclidean and dynamic time warping) between time series. The experimental validation of the method has been done using different datasets, some of them obtained from the UCI repositories. The results are very competitive with the reported in previous works. Moreover, their comprehensibility is better than in other approaches with similar results, since the classifiers are formed by a weighted sequence of literals.