In recent years, many content aware methods for media manipulations have gained popularity. Images and videos are analyzed, segmented, and semantic information is extracted to assist many manipulation algorithms. One of the problems that drew much attention in recent years is media retargeting. Due to the increase in the variety of commonly used display devices, and the prevalent use of mobile devices as available means for media intake, media needs to be adapted to different resolutions and aspect ratios. This problem further increases with the explosion of image and video content on the web. One would like to be able to present a feature film on a small iPod, cellular phone or pocket computer, or show photographs on projected presentation systems.
The goal of this course is to present the basic problem of media retargeting and detail the different methods devised recently to solve it. We will start with a short overview of image and video representation and concentrate on the different view points of media as a discrete entity (pixels, graphs) or as a sampling of a continuous entity (signal). We will then present the common pipeline for resizing, used in both discrete and continuous methods. This includes first extracting some importance or saliency maps from the media, and then using this information while applying the different retargeting operators. We will present several ways to define importance maps that use spatial information in images and also temporal information in video.
We would like to thank all our co-authors for the years of inspiring and fruitful collaborations on the subject of media retargeting. Ariel Shamir’s research is funded by the Israel Science Foundation grant 315/07 and the Israel Ministry of Science grant 3-3421. Olga Sorkine’s research is supported in part by an NYU URCF grant and an NSF award IIS-0905502.