A Survey On "Image Retargeting Quality Assessment: A Backward Registration Approach"

Abstract

This paper demonstrate the result of a recent large-scale subjective study of image retargeting quality on a gathering of images produced by several representative image retargeting methods. The image retargeting operators can be broadly categorized into two types: discrete and continuous approaches. In this paper, interpret the image retargeting in a unified system of resampling system period and forward resampling. This paper creates the impression that the geometric change estimation is a productive approach to clear up the relationship between the images. This paper gives a unified interpretation of image retargeting and shows that the geometric change estimation is an efficient way to clarify the relationship between the original and retargeted images. This paper formulates the geometric change estimation as a backward registration issue with the MRF and gives a reasonable and viable arrangement. Under the geometric change guidance this paper develops a novel ARS metric, which is effective and outperforms other existing techniques on freely accessible datasets. Experimental result is the image fusion.