Hybrid methods for 3D models retrieval based on geometric ratios and statistical information

Two hybrid methods, rank integration and hierarchical methods, are proposed for 3D model retrieval by using features based on geometric ratios and statistical information. Both methods are capable of taking advantage of each single retrieval method very well. A contrast experiment for the hybrid methods, single retrieval methods, and weighted method is conducted. Experimental results show that, the hybrid methods can retrieve more accurate results and have more adaptability for the various categories than using any single methods.