Barycenters of Natural Images -- Constrained Wasserstein Barycenters for Image Morphing
This work addresses the challenge of creating visually appealing image transitions for applications in graphics and media, though it is incremental as it builds on existing Wasserstein barycenter methods.
The paper tackles the problem of image morphing by proposing a constrained Wasserstein barycenter approach that enforces an image prior to ensure natural-looking transitions, achieving smooth, minimal-change, and artifact-free intermediate images.
Image interpolation, or image morphing, refers to a visual transition between two (or more) input images. For such a transition to look visually appealing, its desirable properties are (i) to be smooth; (ii) to apply the minimal required change in the image; and (iii) to seem "real", avoiding unnatural artifacts in each image in the transition. To obtain a smooth and straightforward transition, one may adopt the well-known Wasserstein Barycenter Problem (WBP). While this approach guarantees minimal changes under the Wasserstein metric, the resulting images might seem unnatural. In this work, we propose a novel approach for image morphing that possesses all three desired properties. To this end, we define a constrained variant of the WBP that enforces the intermediate images to satisfy an image prior. We describe an algorithm that solves this problem and demonstrate it using the sparse prior and generative adversarial networks.