This paper describes a novel approach to multi-view 3D objectretrieval. To represent each view of the object, its edge contours areextracted at different levels of scale. Each edge contour, in turn, issegmented by its curvature zero crossing points in a multi-scalefashion. This procedure is carried out using the curvature scale spacetechnique which has been selected for MPEG-7 standardisation. A numberof features are then computed for each segment of each edge contour. Theimage is finally represented by the locations of its segments and thevalues of their associated features. In response to an input query,geometric hashing is first used to find the best locally matchedcandidates for the verification stage where the goal is to measure thedistance between the input query edge contours and the correspondingmodel contours after applying a proper transformation. The measurementis then optimised and used as the match value. The method has beensuccessfully tested on a collection of 3D objects consisting of 15aircraft of different shapes
展开▼
