Sketch-Based Image Retrieval Via Adaptive Weighting, Algorithm And Application

Although text-based approaches have already been used in Content-Based Image Retrieval, but sometimes it is still very hard to represent an image structure precisely by keywords. On the other hand, although serveral search engine provide similar image retrieval function, the user may not have the origin image for the retrieval process. As touch devices become more and more popular these days, it would be an attractive pattern if the image user could draw a sketch and then use it to retrieve relevant images.Sketch-Based Image Retrieval (SBIR) had been studied since 1990s, it is still a challenging problem which how to measure the similarity between a sketch and an image with high precision. In this paper, a novel adaptive weighting method is proposed for the matching process of SBIR. We integrate a cost aggregation step into the matching process, both the neighborhood and multi-scale information are taken into account. To further improve the accuracy and efficiency of the proposed algorithm, a novel edge point selection method is proposed and edge points are selected based on its scale information. We also implement a Sketch-Based Image Retrieval system based on our proposed algorithms. The experiments on the public image datasets show that our method can yield promising results.