%0 Conference Paper
%B 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
%D 2016
%T Unsupervised time-series clustering of distorted and asynchronous temporal patterns.
%A S. Mure
%A T Grenier
%A C. R. G. Guttmann
%A Benoit-Cattin, H.
%K asynchronous temporal patterns
%K brain magnetic resonance
%K categ_st2i
%K clustering algorithm
%K Clustering algorithms
%K DTW associations
%K dynamic time warping
%K Image color analysis
%K image sequences
%K Images et ModÃ¨les
%K k-means
%K k-medoids
%K Labex PRIMES
%K learning step
%K Lesions
%K Measurement
%K multispectral satellite image sequences
%K optimal time series warping
%K pattern clustering
%K reseau_international
%K spatiotemporal filtering technique
%K spatiotemporal mean-shift
%K spatiotemporal phenomena
%K synchronized temporal patterns
%K synthetic data
%K time series
%K Time-series alignment
%K Trajectory
%K trajectory constraint
%K Unsupervised clustering
%K unsupervised clustering technique
%K unsupervised time-series clustering methods
%X Most time-series clustering methods, such as k-means or k-medoids, are initialized by prior knowledge about the number of classes or by a learning step. We propose an unsupervised clustering technique based on spatiotemporal mean-shift and optimal time series warping using dynamic time warping (DTW). Our main contribution consists in combining a spatiotemporal filtering technique, which gathers similar and synchronized temporal patterns in image sequences, with a clustering algorithm that applies a trajectory constraint on the DTW associations, thereby discriminating between similar time-series that are temporally shifted or warped. We assess the method's robustness on synthetic data, and demonstrate its versatility on brain magnetic resonance and multispectral satellite image sequences.

%B 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
%P 1263-1267
%8 March
%G eng
%R 10.1109/ICASSP.2016.7471879