Compressed domain image resizing based on saliency histogram equalisation and seam carving

Abstract

With the increasing popularity of mobile devices and internet services, resizing of compressed images is in high demand. In this study, a compressed domain content-aware image resizing (CAIR) system is proposed. It takes a pair of meshes obtained by using saliency histogram equalisation and seam carving to convert the discrete cosine transform (DCT) blocks of the original image into the DCT blocks of the resized one without the burden of decompression and compression computations. Experimental results show that the proposed system is preferable to other existing CAIR algorithms, especially in terms of running time due mainly to the simplicity of histogram manipulation and dynamic programming.