Supervised shape retrieval based on fusion of multiple feature spaces

Abstract

Shape features are powerful clues for object recognition. In this paper, for improving retrieval accuracy, dissimilarities of contour and region-based shape retrieval methods were used. It is assumed that the fusion of two categories of shape feature spaces causes a considerable improvement in retrieval performance. Fusion of multiple feature spaces can be done in constructing shape description vector and in decision phase. The method proposed in this paper is based on kNN by fusion in calculating of dissimilarity between test and other train samples. Our proposed fused kNN versus fusion of multiple kNNs has better accuracy results in shape classification. The proposed approach has been tested on Chicken Piece dataset. In the experiments, our method demonstrates effective performance compared with other algorithms.