|M.Sc Student||Tal Tlusty Shapiro|
|Subject||Modifying Non-Local Variations Across Multiple Views|
|Department||Department of Electrical Engineering||Supervisors||Professor Zelnik-Manor Lihi|
|Dr. Michaeli Tomer|
|Full Thesis text|
We present an algorithm for modifying small non-local variations between repeating structures and patterns in multiple images of the same scene. The modification is consistent across views, even-though the images could have been photographed from different view points and under different lighting conditions.
Our approach extends a recent algorithm by Dekel et al., which modifies non-local variations (NLV) within a single image. To handle a pair of views, we propose two different formulations that enforce additional constraint on top of those of the original
NLV algorithm. In addition, we show that our two views methods can be easily extended to any number of images of the same scene. For this purpose, we present two ways to guarantee multiple-view consistency by adding an additional term to the two views energy function. We also show an extension to correcting images in which the repeating objects are only semantically similar but differ in their appearance. We show that when modifying each image independently the correspondence between them breaks and the geometric structure of the scene gets distorted. Our approach modifies the views while maintaining correspondence, hence, we succeed in modifying appearance and structure variations consistently.
We demonstrate our methods on a number of challenging examples, photographed in different lighting, scales and view points. We show our extension to multiview is applicable for videos and other sets of images of the same scene.