An official website of the United States government
In the era of generative AI, integrating video generation models into robotics opens new possibilities for the general-purpose robot agent. This paper introduces imitation learning with latent video planning (VILP). We propose a latent video diffusion model to generate predictive robot videos that adhere to temporal consistency to a good degree. Our method is able to generate highly time-aligned videos from multiple views, which is crucial for robot policy learning. Our video generation model is highly time-effcient. For example, it can generate videos from two distinct perspectives, each consisting of six frames with a resolution of 96x160 pixels, at a rate of 5 Hz. In the experiments, we demonstrate that VILP outperforms the existing video generation robot policy across several metrics: training costs, inference speed, temporal consistency of generated videos, and the performance of the policy. We also compared our method with other imitation learning methods. Our fndings indicate that VILP can rely less on extensive high-quality task-specifc robot action data while still maintaining robust performance. In addition, VILP possesses robust capabilities in representing multi-modal action distributions. Our paper provides a practical example of how to effectively integrate video generation models into robot policies, potentially offering insights for related felds and directions. For more details, please refer to our open-source repository https://github.com/ZhengtongXu/VILP.
more »
« less
Frame-Recurrent Video Inpainting by Robust Optical Flow Inference
In this paper, we present a new inpainting framework for recovering missing regions of video frames. Compared with image inpainting, performing this task on video presents new challenges such as how to preserving temporal consistency and spatial details, as well as how to handle arbitrary input video size and length fast and efficiently. Towards this end, we propose a novel deep learning architecture which incorporates ConvLSTM and optical flow for modeling the spatial-temporal consistency in videos. It also saves much computational resource such that our method can handle videos with larger frame size and arbitrary length streamingly in real-time. Furthermore, to generate an accurate optical flow from corrupted frames, we propose a robust flow generation module, where two sources of flows are fed and a flow blending network is trained to fuse them. We conduct extensive experiments to evaluate our method in various scenarios and different datasets, both qualitatively and quantitatively. The experimental results demonstrate the superior of our method compared with the state-of-the-art inpainting approaches.
more »
« less
- Award ID(s):
- 1704309
- PAR ID:
- 10155009
- Date Published:
- Journal Name:
- ArXivorg
- ISSN:
- 2331-8422
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Classifier-free guidance (CFG) is a key technique for improving conditional generation in diffusion models, enabling more accurate control while enhancing sample quality. It is natural to extend this technique to video diffusion, which generates video conditioned on a variable number of context frames, collectively referred to as history. However, we find two key challenges to guiding with variable-length history: architectures that only support fixed-size conditioning, and the empirical observation that CFG-style history dropout performs poorly. To address this, we propose the Diffusion Forcing Transformer (DFoT), a video diffusion architecture and theoretically grounded training objective that jointly enable conditioning on a flexible number of history frames. We then introduce History Guidance, a family of guidance methods uniquely enabled by DFoT. We show that its simplest form, vanilla history guidance, already significantly improves video generation quality and temporal consistency. A more advanced method, history guidance across time and frequency further enhances motion dynamics, enables compositional generalization to out-of-distribution history, and can stably roll out extremely long videos.more » « less
-
Representing videos as linear subspaces on Grassmann manifolds has made great strides in action recognition problems. Recent studies have explored the convenience of discriminant analysis by making use of Grassmann kernels. However, traditional methods rely on the matrix representation of videos based on the temporal dimension and suffer from not considering the two spatial dimensions. To overcome this problem, we keep the natural form of videos by representing video inputs as multidimensional arrays known as tensors and propose a tensor discriminant analysis approach on Grassmannian manifolds. Because matrix algebra does not handle tensor data, we introduce a new Grassmann projection kernel based on the tensor-tensor decomposition and product. Experiments with human action databases show that the proposed method performs well compared with the state-of-the-art algorithms.more » « less
-
Optical flow-based video interpolation is a commonly used method to enrich video details by generating new intermediate frames from existing ones. However, it cannot accurately reproduce the trajectories of irregular and fast-moving objects. To achieve the accurate reconstruction of high-speed scenes, incorporating event information during the interpolation process is an effective method. However, existing methods are not suitable for scenarios where events are sparse. To implement high-quality video interpolation for these scenarios, this study incorporates event information into the interpolation process with three-fold ideas: a) treating the moving target as the foreground and using events to delineate the target area, b) matching foreground to background based on the event information to generate temporal frame between keyframes, and c) generating intermediate frames between keyframe and temporal frame with optical flow interpolation. Empirical studies indicate that owing to its efficient incorporation of event information, the proposed framework outperforms state-of-the-art methods in generating high-quality frames for video interpolation.more » « less
-
The analysis and use of egocentric videos for robotic tasks is made challenging by occlusion due to the hand and the visual mismatch between the human hand and a robot end-effector. In this sense, the human hand presents a nuisance. However, often hands also provide a valuable signal, e.g. the hand pose may suggest what kind of object is being held. In this work, we propose to extract a factored representation of the scene that separates the agent (human hand) and the environment. This alleviates both occlusion and mismatch while preserving the signal, thereby easing the design of models for downstream robotics tasks. At the heart of this factorization is our proposed Video Inpainting via Diffusion Model (VIDM) that leverages both a prior on real-world images (through a large-scale pre-trained diffusion model) and the appearance of the object in earlier frames of the video (through attention). Our experiments demonstrate the effectiveness of VIDM at improving inpainting quality on egocentric videos and the power of our factored representation for numerous tasks: object detection, 3D reconstruction of manipulated objects, and learning of reward functions, policies, and affordances from videos.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/https://arxiv.org/pdf/1905.02882.pdf
