Semi-automated video morphing

Semi-automated video morphing
Jing Liao, Rodolfo Lima, Diego Nehab, Hugues Hoppe, Pedro Sander.
Eurographics Symposium on Rendering, 2014.
Transition across two videos using optimized spatiotemporal alignment.
Abstract: We explore creating smooth transitions between videos of different scenes. As in traditional image morphing, good spatial correspondence is crucial to prevent ghosting, especially at silhouettes. Video morphing presents added challenges. Because motions are often unsynchronized, temporal alignment is also necessary. Applying morphing to individual frames leads to discontinuities, so temporal coherence must be considered. Our approach is to optimize a full spatiotemporal mapping between the two videos. We reduce tedious interaction by letting the optimization derive the fine-scale map given only sparse user-specified constraints. For robustness, the optimization objective examines structural similarity of the video content. We demonstrate the approach on a variety of videos, obtaining results using few explicit correspondences.
Hindsights: No hindsights yet.
Example results: