Proceedings Article•
SparseDTW: a novel approach to speed up dynamic time warping
01 Dec 2009-pp 117-127
TL;DR: SparseDTW as discussed by the authors exploits the existence of similarity and/or correlation between the time series to compute the dynamic time warping distance between two time series that always yields the optimal result.
Abstract: We present a new space-efficient approach, (SparseDTW), to compute the Dynamic Time Warping (DTW) distance between two time series that always yields the optimal result. This is in contrast to other known approaches which typically sacrifice optimality to attain space efficiency. The main idea behind our approach is to dynamically exploit the existence of similarity and/or correlation between the time series. The more the similarity between the time series the less space required to compute the DTW between them. To the best of our knowledge, all other techniques to speedup DTW, impose apriori constraints and do not exploit similarity characteristics that may be present in the data. We conduct experiments and demonstrate that SparseDTW outperforms previous approaches.
Citations
More filters
279 citations
TL;DR: This paper discusses the development of a natural gesture user interface that tracks and recognizes in real time hand gestures based on depth data collected by a Kinect sensor and outperforms most of the solutions for the static recognition of sign digits and is similar in terms of performance for the dynamic recognition of popular signs and for the sign language alphabet.
Abstract: This paper discusses the development of a natural gesture user interface that tracks and recognizes in real time hand gestures based on depth data collected by a Kinect sensor. The interest space corresponding to the hands is first segmented based on the assumption that the hand of the user is the closest object in the scene to the camera. A novel algorithm is proposed to improve the scanning time in order to identify the first pixel on the hand contour within this space. Starting from this pixel, a directional search algorithm allows for the identification of the entire hand contour. The $k$ -curvature algorithm is then employed to locate the fingertips over the contour, and dynamic time warping is used to select gesture candidates and also to recognize gestures by comparing an observed gesture with a series of prerecorded reference gestures. The comparison of results with state-of-the-art approaches shows that the proposed system outperforms most of the solutions for the static recognition of sign digits and is similar in terms of performance for the static and dynamic recognition of popular signs and for the sign language alphabet. The solution simultaneously deals with static and dynamic gestures as well as with multiple hands within the interest space. An average recognition rate of 92.4% is achieved over 55 static and dynamic gestures. Two possible applications of this work are discussed and evaluated: one for interpretation of sign digits and gestures for a friendlier human-machine interaction and the other one for the natural control of a software interface.
219 citations
TL;DR: A LogDet divergence-based metric learning with triplet constraint model which can learn Mahalanobis matrix with high precision and robustness is established.
Abstract: Multivariate time series (MTS) datasets broadly exist in numerous fields, including health care, multimedia, finance, and biometrics. How to classify MTS accurately has become a hot research topic since it is an important element in many computer vision and pattern recognition applications. In this paper, we propose a Mahalanobis distance-based dynamic time warping (DTW) measure for MTS classification. The Mahalanobis distance builds an accurate relationship between each variable and its corresponding category. It is utilized to calculate the local distance between vectors in MTS. Then we use DTW to align those MTS which are out of synchronization or with different lengths. After that, how to learn an accurate Mahalanobis distance function becomes another key problem. This paper establishes a LogDet divergence-based metric learning with triplet constraint model which can learn Mahalanobis matrix with high precision and robustness. Furthermore, the proposed method is applied on nine MTS datasets selected from the University of California, Irvine machine learning repository and Robert T. Olszewski’s homepage, and the results demonstrate the improved performance of the proposed approach.
140 citations
01 Jan 2018
TL;DR: A different type of data-driven approaches allows to perform classification directly on the raw time-series data, avoiding the features’ extraction phase, among these approaches, dynamic time warping and symbolic-based methodologies have been widely applied in many application areas.
Abstract: Chapter Overview The diffusion in power systems of distributed renewable energy resources, electric vehicles, and controllable loads has made advanced monitoring systems fundamental to cope with the consequent disturbances in power flows; advanced monitoring systems can be employed for anomaly detection, root cause analysis, and control purposes. Several machine learning-based approaches have been developed in the past recent years to detect if a power system is running under anomalous conditions and, eventually, to classify such situation with respect to known problems. One of the aspects, which makes Power Systems challenging to be tackled, is that the monitoring has to be performed on streams of data that have a time-series evolution; this issue is generally tackled by performing a features’ extraction procedure before the classification phase. The features’ extraction phase consists of translating the informative content of time-series data into scalar quantities: such procedure may be a time-consuming step that requires the involvement of process experts to avoid loss of information in the making; moreover, extracted features designed to capture certain behaviors of the system, may not be informative under unseen conditions leading to poor monitoring performances. A different type of data-driven approaches, which will be reviewed in this chapter, allows to perform classification directly on the raw time-series data, avoiding the features’ extraction phase: among these approaches, dynamic time warping and symbolic-based methodologies have been widely applied in many application areas. In the following, pros and cons of each approach will be discussed and practical implementation guidelines will be provided.
88 citations
TL;DR: This article presents a longitudinal data partitioning algorithm based on the shapes of the trajectories rather than on classical distances that revealed a “rapid decline” patient group that was not found by the classical methods.
Abstract: Background
Longitudinal data are data in which each variable is measured repeatedly over time. One possibility for the analysis of such data is to cluster them. The majority of clustering methods group together individual that have close trajectories at given time points. These methods group trajectories that are locally close but not necessarily those that have similar shapes. However, in several circumstances, the progress of a phenomenon may be more important than the moment at which it occurs. One would thus like to achieve a partitioning where each group gathers individuals whose trajectories have similar shapes whatever the time lag between them.
Method
In this article, we present a longitudinal data partitioning algorithm based on the shapes of the trajectories rather than on classical distances. Because this algorithm is time consuming, we propose as well two data simplification procedures that make it applicable to high dimensional datasets.
Results
In an application to Alzheimer disease, this algorithm revealed a “rapid decline” patient group that was not found by the classical methods. In another application to the feminine menstrual cycle, the algorithm showed, contrarily to the current literature, that the luteinizing hormone presents two peaks in an important proportion of women (22%).
76 citations
References
More filters
Book•
01 Jan 1993
TL;DR: This book presents a meta-modelling framework for speech recognition that automates the very labor-intensive and therefore time-heavy and therefore expensive and expensive process of manually modeling speech.
Abstract: 1. Fundamentals of Speech Recognition. 2. The Speech Signal: Production, Perception, and Acoustic-Phonetic Characterization. 3. Signal Processing and Analysis Methods for Speech Recognition. 4. Pattern Comparison Techniques. 5. Speech Recognition System Design and Implementation Issues. 6. Theory and Implementation of Hidden Markov Models. 7. Speech Recognition Based on Connected Word Models. 8. Large Vocabulary Continuous Speech Recognition. 9. Task-Oriented Applications of Automatic Speech Recognition.
8,442 citations
NEC1
TL;DR: This paper reports on an optimum dynamic progxamming (DP) based time-normalization algorithm for spoken word recognition, in which the warping function slope is restricted so as to improve discrimination between words in different categories.
Abstract: This paper reports on an optimum dynamic progxamming (DP) based time-normalization algorithm for spoken word recognition. First, a general principle of time-normalization is given using time-warping function. Then, two time-normalized distance definitions, called symmetric and asymmetric forms, are derived from the principle. These two forms are compared with each other through theoretical discussions and experimental studies. The symmetric form algorithm superiority is established. A new technique, called slope constraint, is successfully introduced, in which the warping function slope is restricted so as to improve discrimination between words in different categories. The effective slope constraint characteristic is qualitatively analyzed, and the optimum slope constraint condition is determined through experiments. The optimized algorithm is then extensively subjected to experimental comparison with various DP-algorithms, previously applied to spoken word recognition by different research groups. The experiment shows that the present algorithm gives no more than about two-thirds errors, even compared to the best conventional algorithm.
5,906 citations
Proceedings Article•
31 Jul 1994TL;DR: Preliminary experiments with a dynamic programming approach to pattern detection in databases, based on the dynamic time warping technique used in the speech recognition field, are described.
Abstract: Knowledge discovery in databases presents many interesting challenges within the context of providing computer tools for exploring large data archives. Electronic data repositories are growing quickly and contain data from commercial, scientific, and other domains. Much of this data is inherently temporal, such as stock prices or NASA telemetry data. Detecting patterns in such data streams or time series is an important knowledge discovery task. This paper describes some preliminary experiments with a dynamic programming approach to the problem. The pattern detection algorithm is based on the dynamic time warping technique used in the speech recognition field.
3,229 citations
TL;DR: This work introduces a novel technique for the exact indexing of Dynamic time warping and proves its vast superiority over all competing approaches in the largest and most comprehensive set of time series indexing experiments ever undertaken.
Abstract: The problem of indexing time series has attracted much interest. Most algorithms used to index time series utilize the Euclidean distance or some variation thereof. However, it has been forcefully shown that the Euclidean distance is a very brittle distance measure. Dynamic time warping (DTW) is a much more robust distance measure for time series, allowing similar shapes to match even if they are out of phase in the time axis. Because of this flexibility, DTW is widely used in science, medicine, industry and finance. Unfortunately, however, DTW does not obey the triangular inequality and thus has resisted attempts at exact indexing. Instead, many researchers have introduced approximate indexing techniques or abandoned the idea of indexing and concentrated on speeding up sequential searches. In this work, we introduce a novel technique for the exact indexing of DTW. We prove that our method guarantees no false dismissals and we demonstrate its vast superiority over all competing approaches in the largest and most comprehensive set of time series indexing experiments ever undertaken.
1,925 citations
TL;DR: A computer system is described in which isolated words, spoken by a designated talker, are recognized through calculation of a minimum prediction residual through optimally registering the reference LPC onto the input autocorrelation coefficients using the dynamic programming algorithm.
Abstract: A computer system is described in which isolated words, spoken by a designated talker, are recognized through calculation of a minimum prediction residual. A reference pattern for each word to be recognized is stored as a time pattern of linear prediction coefficients (LPC). The total log prediction residual of an input signal is minimized by optimally registering the reference LPC onto the input autocorrelation coefficients using the dynamic programming algorithm (DP). The input signal is recognized as the reference word which produces the minimum prediction residual. A sequential decision procedure is used to reduce the amount of computation in DP. A frequency normalization with respect to the long-time spectral distribution is used to reduce effects of variations in the frequency response of telephone connections. The system has been implemented on a DDP-516 computer for the 200-word recognition experiment. The recognition rate for a designated male talker is 97.3 percent for telephone input, and the recognition time is about 22 times real time.
1,588 citations