Fuzzy Contour Matching for 3D Reconstruction and Retrieval

An approach for 3D reconstruction and partial retrieval from 2D image stacks using contour tree structure and fuzzy contour matching model is proposed. We use a contour-based image segmentation method to extract object contours from 2D image slices and convert the original image stack into a standard XML database. A fuzzy contour matching model is designed to group the related contour objects on different image slices to form 3D components. The proposed methods have been implemented for various parts of Crayfish neuron represented in 2D confocal images. The experimental results suggest the suitability of the system for individual neuron branch search in 3D image stacks.