Reasoning with spatial relations over high-content images

Spatial relation and configuration modeling issues are gaining momentum in image analysis and pattern recognition fields in the perspective of mining high-content images or large scale image databases in a more expressive way than purely statistically. Continuing our previous efforts whereby we developed specific efficient morphological tools performing on mesh representation like Delaunay triangulations, we propose to formalize spatial relation modeling techniques dedicated to unorganized point sets. We provide an original mesh lattice framework more convenient for structural representations of large amount of image data by the means of interest points sets and their morphological analysis. The set of designed numerical operators is based on a specific dilation operator making it possible the representation of concepts like “between” or “left of” over sparse representations such as graphs. Then, for the sake of illustration and discussion, we apply these new tools to high-level queries in microscopic histo-pathological images and structural analysis of macroscopic images.