Topological Features Extraction Based On Vision

Topological invariance is the main content of the topology, as found in Chen Lin of the psychological experiment, human visual perception is starting from the topological properties of the object.To learn the handling characteristics of the Visual system for computer,combining visual cognition and the theory of topological spaces for identify the target quickly.This research works in the thesis are:First of all, Exploring the mechanism of the human eye processes Visual information from a psychological perspective, as the reserves of relevant knowledge.Secondly,the connectivity in topological as the primary research. three-dimensional contour extraction using ICP algorithm in combination with snake,in the case of topological properties remains unchanged,dimension projection to the two-dimensional plane.The point se topology analysis in the plane to obtain the connectivity.We use physical validated the feasibility of this derivation,proof the kettle and tries can be homeomorphic transformation.Thirdly,made a study of three dimensional model of connectivity. Because the point clouds acquired information is distributed, point in a three dimensional space structure needs a little cloud matching, surface reconstruction, such as pretreatment,but Digital topology is the topology branch,topological invariance analysis by genus formula on voxel element directly,non-connectivity.Finally, in the acquisition of information on the image, find a way to combine pictures and topological knowledge base the similarity measure, different topologies are simulated using Matlab histogram, Comparative histogram object for topological classification.