Modifying Non-Local Variations

Across Multiple Views

*Tal Tlusty     *Tomer Michaeli     +Tali Dekel     *Lihi Zelnik-Manor    


*Technion – Israel Institute of Technology

+Google Research

Gaudi’s Casa Mila is famous for its irregular shapes. Single-image NLV idealizes each image independently, resulting in utterly different geometries – the building’s facade is flat and some of the porches are straightened only in the top image. Reconstruction of a corresponding 3D structure completely fails. Conversely, our multiview NLV approach reduces the irregularities in both images in a synchronized manner and the corresponding 3D shape shows straightened verandas.



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.



“Modifying Non-Local Variations Across Multiple Views” [pdf] [Supplementary] [Results] [K-Views Extension]