Novel image retrieval algorithm using salient closed boundary
Zhou Li-hua · Systems engineering and electronics · 2007
Having been analyzed the drawbacks of image retrieval methods based on edges,a novel image retrieval algorithm based on salient closed boundary is presented.Firstly,the ratio contour,a novel graph-based method is performed to extract salient closed boundaries from noisy images.Secondly,salient closed boundaries are selected and each salient closed boundary is described by the feature: edge angular histogram,the image is represented by the feature of salient closed boundaries.Finally,the similarities are measured by the angular histogram of the salient closed boundary.Because the salient closed boundary represent the prominent structure of image and eliminate the noise of the short edges,the algorithm is very effective for image retrieval.Experimental results demonstrate that the proposed method is more effective than the traditional algorithm.