Skip to main content
main-content

Tipp

Weitere Artikel dieser Ausgabe durch Wischen aufrufen

Erschienen in: Neural Processing Letters 3/2022

21.01.2022

Non-Interlaced Dynamic Time Warping for Distance Between Matrixes

verfasst von: Junjie Li, Cuifang Gao, Ping Yin

Erschienen in: Neural Processing Letters | Ausgabe 3/2022

Einloggen, um Zugang zu erhalten
share
TEILEN

Abstract

DTW measures is an indication of dissimilarity, which has a wide range of applications in pattern recognition and machine learning. Nevertheless, it is difficult to directly calculate the distance between unequal-sized matrices, which can potentially be solved by using the proposed Non-Interlaced Dynamic Time Warping (Non-Interlaced DTW) algorithm in this paper. In the new algorithm, matrices are separated into row vectors, and DTW distance matrix is set up based on the distances between row vectors. By introducing dynamic programming, the minimum cumulative distance and shortest path of the DTW distance matrix can be achieved, as well as the optimal match and the rows-alignment between row vectors. Non-Interlaced DTW algorithm defines the normalized cumulative distance of DTW distance matrix as the final distance measure. The performance of the algorithm is tested and verified by experiments on MNIST database, FVC2004 dataset and Caltech database. The results illustrate that the Non-Interlaced DTW can get accurate similarity between unequal-sized matrixes. By directly obtaining the distance between matrices, Non-Interlaced DTW provides a new idea of similarity measure for unequal-sized matrices and has the promising application when dealing with high-dimensional data.
Literatur
1.
Zurück zum Zitat Abramson N, Braverman D, Sebestyen G (1963) Pattern recognition and machine learning. IEEE Trans Inform Theory 9(4):257–261 CrossRef Abramson N, Braverman D, Sebestyen G (1963) Pattern recognition and machine learning. IEEE Trans Inform Theory 9(4):257–261 CrossRef
2.
Zurück zum Zitat Yang J, Ma Y, Zhang X, Li S, Zhang Y (2017) An initialization method based on hybrid distance for k-means algorithm. Neural Comput 29(11):3094–3117 MathSciNetCrossRef Yang J, Ma Y, Zhang X, Li S, Zhang Y (2017) An initialization method based on hybrid distance for k-means algorithm. Neural Comput 29(11):3094–3117 MathSciNetCrossRef
3.
Zurück zum Zitat Alpaydin E (2016) Introduction to machine learning, 3rd edn. China Machine Press, Beijing MATH Alpaydin E (2016) Introduction to machine learning, 3rd edn. China Machine Press, Beijing MATH
4.
Zurück zum Zitat Górecki T, Łuczak M (2014) Non-isometric transforms in time series classification using DTW. Knowl-Based Syst 61(2):98–108 CrossRef Górecki T, Łuczak M (2014) Non-isometric transforms in time series classification using DTW. Knowl-Based Syst 61(2):98–108 CrossRef
5.
Zurück zum Zitat Krzeszowski T et al (2014) DTW-based gait recognition from recovered 3-d joint angles and inter-ankle distance. In: International conference on computer vision and graphics (ICCVG), pp 356–363 Krzeszowski T et al (2014) DTW-based gait recognition from recovered 3-d joint angles and inter-ankle distance. In: International conference on computer vision and graphics (ICCVG), pp 356–363
6.
Zurück zum Zitat Okawa M (2020) Time-series averaging and local stability-weighted dynamic time warping for online signature verification. Pattern Recogn 112:107699 CrossRef Okawa M (2020) Time-series averaging and local stability-weighted dynamic time warping for online signature verification. Pattern Recogn 112:107699 CrossRef
8.
Zurück zum Zitat Switonski A, Krzeszowski T, Josinski H, Kwolek B, Wojciechowski K (2018) Gait recognition on the basis of markerless motion tracking and DTW transform. BIOMS Iet 7(5):415–422 CrossRef Switonski A, Krzeszowski T, Josinski H, Kwolek B, Wojciechowski K (2018) Gait recognition on the basis of markerless motion tracking and DTW transform. BIOMS Iet 7(5):415–422 CrossRef
9.
Zurück zum Zitat Keogh E J, Pazzani M J (2001) Derivative dynamic time warping. In: First SIAM international conference on data mining, pp 1–11 Keogh E J, Pazzani M J (2001) Derivative dynamic time warping. In: First SIAM international conference on data mining, pp 1–11
10.
Zurück zum Zitat Ratanamahatana CA, Keogh EJ (2008) Making time-series classification more accurate using learned constraints. In: Proceedings of the fourth SIAM international conference on data mining, pp 11–22 Ratanamahatana CA, Keogh EJ (2008) Making time-series classification more accurate using learned constraints. In: Proceedings of the fourth SIAM international conference on data mining, pp 11–22
11.
Zurück zum Zitat Jeong YS, Jeong MK, Omitaomu OA (2011) Weighted dynamic time warping for time series classification. Pattern Recogn 44(9):2231–2240 CrossRef Jeong YS, Jeong MK, Omitaomu OA (2011) Weighted dynamic time warping for time series classification. Pattern Recogn 44(9):2231–2240 CrossRef
12.
Zurück zum Zitat Zhao J, Itti L (2018) shapeDTW: shape dynamic time warping. Pattern Recogn 74:171–184 CrossRef Zhao J, Itti L (2018) shapeDTW: shape dynamic time warping. Pattern Recogn 74:171–184 CrossRef
13.
Zurück zum Zitat Li H et al (2020) Adaptively constrained dynamic time warping for time series classification and clustering. Inf Sci 534:97–116 MathSciNetCrossRef Li H et al (2020) Adaptively constrained dynamic time warping for time series classification and clustering. Inf Sci 534:97–116 MathSciNetCrossRef
14.
Zurück zum Zitat Zhang QS, Zhu SC (2018) Visual interpretability for deep learning: a survey. Front Inf Technol Electron Eng 19(01):27–39 CrossRef Zhang QS, Zhu SC (2018) Visual interpretability for deep learning: a survey. Front Inf Technol Electron Eng 19(01):27–39 CrossRef
16.
Zurück zum Zitat Yellamraju T, Boutin M (2018) Clusterability and clustering of images and other “real” high-dimensional data. IEEE Trans Image Process 27(4):1927–1938 MathSciNetCrossRef Yellamraju T, Boutin M (2018) Clusterability and clustering of images and other “real” high-dimensional data. IEEE Trans Image Process 27(4):1927–1938 MathSciNetCrossRef
17.
Zurück zum Zitat Wang L, Zhang Y, Feng J (2005) On the Euclidean distance of images. IEEE Trans Pattern Anal Mach Intell 27(8):1334–1339 CrossRef Wang L, Zhang Y, Feng J (2005) On the Euclidean distance of images. IEEE Trans Pattern Anal Mach Intell 27(8):1334–1339 CrossRef
18.
19.
Zurück zum Zitat Maio D et al (2004) FVC2004: third fingerprint verification competition. Lect Notes Comput Sci 3072:1–7 CrossRef Maio D et al (2004) FVC2004: third fingerprint verification competition. Lect Notes Comput Sci 3072:1–7 CrossRef
20.
Zurück zum Zitat Gonzalez R, Woods R (2017) Digital image processing, 3rd edn. Publishing House of Electronics Industry, Beijing Gonzalez R, Woods R (2017) Digital image processing, 3rd edn. Publishing House of Electronics Industry, Beijing
21.
Zurück zum Zitat Deng L (2012) The MNIST database of handwritten digit images for machine learning research [Best of the Web]. IEEE Signal Process Mag 29(6):141–142 CrossRef Deng L (2012) The MNIST database of handwritten digit images for machine learning research [Best of the Web]. IEEE Signal Process Mag 29(6):141–142 CrossRef
22.
Zurück zum Zitat Feifei L, Fergus R, Perona P (2006) One-shot learning of object categories. IEEE Trans Pattern Anal Mach Intell 28(4):594–611 CrossRef Feifei L, Fergus R, Perona P (2006) One-shot learning of object categories. IEEE Trans Pattern Anal Mach Intell 28(4):594–611 CrossRef
23.
Zurück zum Zitat Yan J, Li C, Li Y, Cao G (2018) Adaptive discrete hypergraph matching. IEEE Trans Cybern 48(2):765–779 CrossRef Yan J, Li C, Li Y, Cao G (2018) Adaptive discrete hypergraph matching. IEEE Trans Cybern 48(2):765–779 CrossRef
Metadaten
Titel
Non-Interlaced Dynamic Time Warping for Distance Between Matrixes
verfasst von
Junjie Li
Cuifang Gao
Ping Yin
Publikationsdatum
21.01.2022
Verlag
Springer US
Erschienen in
Neural Processing Letters / Ausgabe 3/2022
Print ISSN: 1370-4621
Elektronische ISSN: 1573-773X
DOI
https://doi.org/10.1007/s11063-022-10739-3

Weitere Artikel der Ausgabe 3/2022

Neural Processing Letters 3/2022 Zur Ausgabe