A robust two-stage shape retrieval (TSR) method is proposed to address the 2Dshape retrieval problem. Most state-of-the-art shape retrieval methods arebased on local features matching and ranking. Their retrieval performance isnot robust since they may retrieve globally dissimilar shapes in high ranks. Toovercome this challenge, we decompose the decision process into two stages. Inthe first irrelevant cluster filtering (ICF) stage, we consider both global andlocal features and use them to predict the relevance of gallery shapes withrespect to the query. Irrelevant shapes are removed from the candidate shapeset. After that, a local-features-based matching and ranking (LMR) methodfollows in the second stage. We apply the proposed TSR system to MPEG-7,Kimia99 and Tari1000 three datasets and show that it outperforms all otherexisting methods. The robust retrieval performance of the TSR system isdemonstrated.
展开▼
