A 3D model retrieval system based on the multi-resolution frequency descriptors

With the recent advancement in computer graphic, 3D models have become widely used in computer aided design, computer animations, 3D printer, digital library, and so on. The searching for interested 3D models becomes an essential issue. Techniques for effective and efficient content-based 3D model retrieval have therefore become an important research topic. In this paper, five transforms： 3D Discrete Fourier Transform (3D-DFT), 3D Angular Radial Transform (3D-ART), 3D Discrete Wavelet Transform (3D-DWT), 3D Discrete Wavelet Packet Transform (3D-DWPT), and Spherical Harmonic Transform (SHT) are used to extract the multi-resolution frequency descriptors for 3D model retrieval. Experimental results show the compared performances of different frequency descriptors on Princeton Shape Benchmark database (PSB).