3D Content-based Search based on 3D Krawtchouk moments

Abstract

In this paper a novel method for 3D content-based search and retrieval is proposed. Guided by the imperative need for a reliable 3D content based search tool and the very interesting results of research work done in the past on the performance of Krawtchouk moments and Krawtchouk moment invariants in image processing, Weighted 3D Krawtchouk moments are introduced for efficient 3D analysis which are suitable for content-based search and retrieval applications. The proposed method was tested on Princeton Shape Benchmark. Experiments have shown that the proposed method is superior in terms of precision-recall comparing with other well-known methods reported in the literature. View