A wavelet method for the characterization of spatiotemporal patterns

This paper introduces a wavelet-based method for the characterization of spatiotemporal patterns. Based on the wavelet multiresolution analysis, two wavelet indices, multiscale accumulative density (MAD) and multiscale accumulative change (MAC), are proposed for the characterization of the dynamics of the spatiotemporal patterns. Both indices are constructed by using orthogonal wavelet projection operators. The MAD is a measure of the spatial complexity of a pattern at a given time, whereas the MAC characterizes the spatial complexity of instantaneous change of the spatiotemporal patterns at a given time. The ratio of the MAD indices between the lowest and the highest scales reflects the order of coherence in a pattern. The time series of both MAD and MAC provide the dynamical information of morphological pattern evolutions. Numerical experiments based on the Cahn–Hilliard equation indicate that the proposed method is efficient for quantitatively characterizing the dynamics of the spatiotemporal patterns.