Speaker: Tal Tlusty
Affiliation: Dept. of Electrical Engineering, Technion
Modifying Non-Local Variations Across Multiple Views
Repetition of patterns and structures is a widespread phenomenon, e.g, in the leaves of plants and flowers, animal furs or rocks and sand dunes. Recurring structures can be also found in man-made environments, for example, a row of chairs in a large stadium or a pile of boxes in a shoe store. In many cases, the recurring structures are not perfectly identical and sometimes the deviations from an‘ideal’ structure are small and hard to notice by the naked eye. Revealing these deviations may be useful in many situations, for example, revealing deformations in a production line or detecting irregular cells in a petri dish. Modifying these variations and correcting them could be useful for beautification of images. In the modern media age we live in, oftentimes multiple images of the same scene are taken. Different people capture the same monuments, events, and objects, and share them publicly.
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. 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.