Model Retrieval Based on Flattened Shape Comparison in 3D Databases

A method that flatten surface of 3D model is presented. It can flatten the surface of model into a 2D shape as well as automatically cut and partition. The method avoids global and local self-intersections and minimizes the total length of the introduced seams. Then, a 3D models retrieval methodology based on the shape comparison is proposed. By using the flattened surface, pole mapped version normalization and Fourier transform, the shape similarity comparison is performed. The method is effective in capturing characteristics of surface shape. Finally, the experiment results show that our method is better than other methods.