Object-Coherence Warping for Stereoscopic Image Retargeting

• Main Authors: Shu-HuaiChang, 張述淮
• Other Authors: Tong-Yee Lee
• Format: Others
• Language: en_US
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/92805539722182930958

碩士 === 國立成功大學 === 資訊工程學系碩博士班 === 101 === This paper addresses the topic of content-aware stereoscopic image retargeting. The key to this topic is consistently adapting a stereoscopic image to fit displays with various aspect ratios and sizes while preserving the shape and disparity of visually salient content. Most methods focus on preserving salient content through nonlinear image warping in which distortions caused by warping are propagated to homogenous and low-significance regions. However, disregarding the consistency of object deformation sometimes results in apparent distortions in both the shapes and disparities of objects. An object-coherence warping scheme with an object-based significance measurement is proposed in this paper to reduce this unwanted distortion. The basic idea is utilizing information of matched objects rather than matched pixels in warping. Such information implies object correspondences in a stereoscopic image pair, which allows the generation of object significance map and the consistent preservation of objects. This strategy enables our method to consistently preserve not only the spatial shapes but also the disparities of visually salient objects, yielding a pleasing content-aware retargeting. Qualitative and quantitative analyses of various stereoscopic images show that our results are better than those generated by related methods in terms of consistency of object preservation, especially for images with structure lines and evident foreground objects.