D-MASC: A Novel Search Strategy for Detecting Regions of Interest in Linear Parameter Space

摘要

The parameter space transform has been utilized over decades in context of edge detection in the computer vision domain. However the usage of the parameter space transform in context of clustering is a more recent application with the purpose of detecting (hyper)linear correlated clusters. The runtime for detecting edges or hyperlinear correlations can be very high. The contribution of our work is to provide a novel search strategy in order to accelerate the detection of regions of interest in parameter space serving as a foundation for faster detection of edges and linear correlated clusters.