Progressive Image Segmentation by Wavelet Transform for Content-Based Image Retrieval

Abstract

Recently, image segmentation is widely used in the multimedia applications. For a content-based image retrieval system, appropriate image segmentation techniques lead the image retrieval to be performed effectively and vitally. In this paper, we propose a novel segmentation algorithm based on progressive wavelet transform to segment out the notable objects in an image. The proposed algorithm, which is fitting for the use of the simplest wavelet transform, i.e., Haar wavelet transform, has high efficiency and feasibility. To further handle the stress of color differences, we invoke the whitening transform for the composition of processed color signals to meet the visual perception. The progressive processing will recursively remove the insignificant color differences by the suppression thresholds resulting from the proposed assignment algorithm. Thus, the colors within image objects will be unified level-by-level. Consequently, our algorithm can not only effectively extract significant objects, but also make the decomposition of the Haar transform be automatically stopped according to picture characteristics. Further, our method is very efficient for the segment that its speed is almost four times as fast as the previous Lin's method in average.