3D Model Retrieval Based on Weighted Matching Point

As collections of 3D content grown rapidly, the ability to perform fast and accurate retrieval from a database of models or the Internet has become a necessity. This paper proposed a weighted matching point retrieval method of the 3D model according to the characteristics of the contour images of projected 3D models. This algorithm extracts SIFT features from the 2D projections of the 3D model, and then using the weighted parameters trained with SIFT to measure the similarity of 3D models. By comparing with the performance of ray-based method and D2 shape distribution method and Zernike moment method on the PSB, the experiment results show that the method is stable, and has better retrieval performance.