Backward registration based objective quality assessment for image re targeting

Backward registration based objective quality assessment for image re targeting

Backward registration based objective quality assessment for image re targeting. This paper show the result of a recent large-scale subjective study of image retargeting quality on a collection of images produced by several representative image retargeting techniques. In this paper, we interpret the image retargeting in a unified framework of resampling network era and forward resampling. We appear that the geometric change estimation is an efficient way to clear up the relationship between the images.

Image processing

The typical discrete approaches include the manual Cropping (CR), Seam-Carving (SC), Shift-Map(SM) and Multi-Operator (MO), while uniform Scaling in one dimension (SCL), non-homogeneous Warping (WARP), Streaming Video (SV) and Scale-and-Stretch (SNS) are representative continuous methods. Both kinds of content-aware approaches remove pixels or warp the image to the targeted size according to the visual importance of image contents.

An efficient objective quality assessment for image retargeting would thus be useful to select the best result from a pool of retargeted results given a single input image, without the need to rely on costly user studies.

The typical discrete approaches mainly include manual Cropping (CR), Seam-Carving (SC), and Shift-Map (SM), while uniform Scaling in one dimension (SCL), nonhomogeneous Warping (WARP), Streaming Video (SV) and Scale-and-Stretch (SNS) are representative continuous methods. Both kinds of content-aware approaches remove pixels or warp the image to the targeted resolution according to the visual importance of image content. The purpose of the content-aware image retargeting is to preserve visually important content and structure (i.e. to reduce information loss), and at the same time, to limit the visual distortion in the retargeted image.

 

 

Leave a Reply

Your email address will not be published. Required fields are marked *