A VISUAL SIMILARITY- BASED 3D SEARCH ENGINE

Retrieval systems for 3D objects are required because 3D databases used around the web are growing. In this paper, we propose a visual similarity based search engine for 3D objects. The system is based on a new representation of 3D objects given by a 3D closed curve that captures all information about the surface of the 3D object. We propose a new 3D descriptor, which is a combination of three signatures of this new representation, and we implement it in our interactive web based search engine. Our method is compared to some state of the art methods, tested using the Princeton-Shape Benchmark as a large database of 3D objects. The experimental results show that the enhanced curve analysis descriptor performs well.