Objective image quality assessment using perceptual distortion for image retargeting

Image quality assessment of retargeted images is a challenging problem and used to evaluate the quality of image accurately and automatically. In this paper, a full reference image quality metric for assessing visual quality of retargeted image is presented using deformable spatial pyramid (DSP) algorithm. The metric measures geometric and information loss distortion based on fast dense corresponding scheme. Experimental results using NRID database are obtained to measure the performance of the metric and it is found that DSP algorithm outperforms SIFT in terms of execution speed and accuracy.