In this paper, we demonstrate a fully automatic method for converting a still image into a realistic animated looping video.

Resultant videos, codes, and datasets will be available at. We demonstrate the effectiveness of our method through comparisons with the state-of-the-arts on video prediction as well as appearance manipulation. The output frames can be looped like cinemagraph, and also be controlled directly by specifying latent codes or indirectly via visual annotations. This design is crucial to alleviate error accumulation in long-term predictions, which is the essential problem in previous recurrent approaches. These intermediate data are applied not to each previous output frame, but to the input image only once for each output frame. Unlike previous methods that infer output frames directly, our CNNs predict spatially-smooth intermediate data, i.e., for motion, flow fields for warping, and for appearance, color transfer maps, via self-supervised learning, i.e., without explicitly-provided ground truth. We thus learn them separately and predict them with decoupled control while handling future uncertainty in both predictions by introducing latent codes. Our key observation is that the motion (e.g., moving clouds) and appearance (e.g., time-varying colors in the sky) in natural scenes have different time scales. This paper proposes a method that can create a high-resolution, long-term animation using convolutional neural networks (CNNs) from a single landscape image where we mainly focus on skies and waters. Our study demonstrates that our system allows users to achieve better results while requiring little effort compared to existing solutions for the 3D Ken Burns effect creation.Īutomatic generation of a high-quality video from a single image remains a challenging task despite the recent advances in deep generative models. Experiments with a wide variety of image content show that our method enables realistic synthesis results. To address disocclusions while maintaining geometrically and temporally coherent synthesis results, we utilize context-aware color- and depth-inpainting to fill in the missing information in the extreme views of the camera path, thus extending the scene geometry of the point cloud. According to this depth estimate, our framework then maps the input image to a point cloud and synthesizes the resulting video frames by rendering the point cloud from the corresponding camera positions. To address the limitations of existing depth estimation methods such as geometric distortions, semantic distortions, and inaccurate depth boundaries, we develop a semantic-aware neural network for depth prediction, couple its estimate with a segmentation-based depth adjustment process, and employ a refinement neural network that facilitates accurate depth predictions at object boundaries. Our framework first leverages a depth prediction pipeline, which estimates scene depth that is suitable for view synthesis tasks. In this paper, we introduce a framework that synthesizes the 3D Ken Burns effect from a single image, supporting both a fully automatic mode and an interactive mode with the user controlling the camera. Existing automatic methods, however, require multiple input images from varying viewpoints. Creating such effects manually is time-consuming and demands sophisticated editing skills. Adding parallax, which results in the 3D Ken Burns effect, enables significantly more compelling results. The Ken Burns effect allows animating still images with a virtual camera scan and zoom. We demonstrate that our system allows the user to create a variety of animations of water scenes. Finally, the user optionally applies a texture synthesis algorithm to recover the appearance of the input image. After selecting the animation for all the regions, the resulting animation is completed. Our system then allows the user to interactively control the speed of the desired animation and select the appropriate animation. Given an input image, alpha masks specifying regions of interest, and sketches specifying flow directions, our system first retrieves appropriate video candidates from the database and create the candidate animations for each region of interest as the composite of the input image and the retrieved videos: this process spends less than one minute by taking advantage of parallel distributed processing. Our method relies on a database of videos of water scenes and video retrieval technique. We present a system to quickly and easily create an animation of water scenes in a single image.

Cinemagraph pro 2.2 license code