A parallel algorithm for structure detection based on wavelet and segmentation analysis

We present a parallel algorithm which allows to recognize rapidly structures in a 3-dimensional set of discrete data points resulting from numerical experiments, and to study their morphological properties. The algorithm consists of two main steps: (1) wavelet analysis in order to separate those data points which belong to structures from uniformly distributed background points, and (2) segmentation analysis in order to label individual structures and their corresponding data points. Parameters which characterize the morphology of these structures may then be extracted easily. The fast parallel implementation on a Connection Machine CM-200 makes the algorithm interesting for other areas in computational physics which require a method for morphological comparisons. The algorithm is illustrated by an example in the field of cosmology for studying the formation of the Large Scale Structure in the Universe. This analysis allows to distinguish clearly qualitatively as well as quantitatively between two models which respectively favour filamentary or clustered structures.