Visual Investigation Using Circular Partitioning of Abstract Images

This paper presents a novel approach for sketch-based image retrieval based on low-level features. The approach enables measuring the similarity be- tween a full color image and a simple black and white sketched query and needs no cost intensive image segmentation. The proposed method can cope with im- ages containing several complex objects in an inhomogeneous background. Two abstract images are obtained using strong edges of the model image and thinned outline of the sketched image. Circular-spatial distribution of pixels in the ab- stract images is used to extract new compact and effective features. The extracted features are scale and rotation invariant and tolerate small translations. The ma- jor contribution of the paper is in rotation invariance property of the proposed approach. A collection of paintings and sketches (ART BANK) is used for test- ing the proposed method. The results are compared with three other well-known approaches within the literature. Experimental results show signi£cant improve- ment in the Recall ratio using the proposed features.