A Spatial Feature for Content-based Image Retrieval
Yi Liu · Computer Engineering and Applications Journal · 2006
Many approaches for extracting color feature are widely used for content-based image retrieval.Each of them has its unique advantages.However,we cannot present the visual characterization of an image comprehensively by simply using a color feature extracting method.In this paper,we present an algorithm to collect the spatial feature of an image.This algorithm creates two graphs through two line scans on the image,one of which is in horizontal direction and the other is in vertical.A system is implemented to test the efficiency of this new spatial feature extracting algorithm.By combining a popular color feature approach,called color auto-correlogram,we demonstrate that this new algorithm can associate with color feature well to enhance the accuracy of image indexing.