The paper demonstrates that large text-to-video diffusion models contain implicit 3D motion information in their intermediate denoising states. By introducing parasitic co-denoising, the authors extract motion directly from the frozen model using a Parasitic Motion Decoder (PMD) that shares the noise schedule and fuses intermediate features. This approach yields efficient, aligned 3D motion generation without training a separate motion model, achieving comparable results to dedicated generators with far fewer parameters.
arXiv:2609.37004v1 Announce Type: new Abstract: We present World2Motion, a framework that generates scene-aware 3D human motion and corresponding video from a single image and a text prompt. While ex...
arXiv:2607. 20628v1 Announce Type: cross Abstract: Real-world video deblurring remains challenging due to diverse motion patterns, complex degradations, and the scarcity of realistic training data, yet robust restoration is critical for downstream pipelines such as mobile imaging and 3D reconstruction.
The paper introduces VoRTeC, a video compression framework that leverages a foundational flow model to encode latent video representations compactly and predict their positions along flow trajectories. By integrating multi‑scale priors and avoiding access to flow‑matching network parameters, VoRTeC achieves one‑step decoding with high perceptual fidelity, while maintaining temporal consistency through tail‑frame reuse and prior caching. Experiments show a 58% reduction in bit consumption compared to prior diffusion‑based methods and a decoding speed increase ranging from 3 to 197 times, reaching 13 FPS at 720p and 32 FPS at 480p.
arXiv:2608. 19556v1 Announce Type: cross Abstract: Streaming autoregressive diffusion models enable real-time, long-horizon video generation, but their training objectives optimize local frame prediction rather than the geometry and dynamics of a coherent world: long rollouts accumulate geometric drift and degrade into static or unnatural motion.
arXiv:2608. 05237v1 Announce Type: cross Abstract: Current few-step autoregressive video diffusion models depend on previous fully denoised clean frames as context for all denoising steps of the current frame.
Forge4D is a feed‑forward model that reconstructs temporally aligned 4D human representations from uncalibrated sparse‑view videos, enabling both novel view and novel time synthesis. It achieves this by jointly streaming 3D Gaussian reconstruction with dense motion prediction, using learnable state tokens for temporal consistency and a self‑supervised retargeting loss for motion prediction. Extensive experiments confirm its effectiveness on in‑domain and out‑of‑domain datasets.
Streaming autoregressive diffusion models enable real-time, long-horizon video generation, but their training objectives optimize local frame prediction rather than the geometry and dynamics of a coherent world: long rollouts accumulate geometric drift and degrade into static or unnatural motion. Recent bidirectional approaches address this problem using rewards signals built upon 3D Gaussian-Splatting reconstruction.
arXiv:2607. 29180v1 Announce Type: cross Abstract: Text-to-motion generation must produce motions that are semantically correct, temporally coherent, and physically plausible.
arXiv:2608. 14706v1 Announce Type: cross Abstract: Standard autoregressive video generation algorithms based on Diffusion and Flow Matching rely on rigid training objectives and static sampling schedules, limiting inference procedures from adapting to the data.
arXiv:2610.07958v1 Announce Type: new Abstract: Feed-forward 3D Gaussian Splatting (3DGS) reconstructs a scene in a single forward pass, replacing per-scene optimization with a network trained across...
The paper compares diffusion and rectified flow objectives within the MotionGPT3 framework for text-driven motion generation. Experiments on HumanML3D show that rectified flow converges faster, achieves strong test performance earlier, and matches or exceeds diffusion quality while requiring fewer sampling steps. The study isolates the generative objective’s impact, demonstrating that rectified flow’s benefits transfer to continuous-latent motion generation.
FilterFusion is a lightweight video denoising framework that blends temporally-aware learned features with a small bank of classical denoising filters. It uses a low-cost guidance signal to adaptively mix filter responses at the original resolution, achieving high quality on real sensor noise datasets while keeping the model to only 109K parameters. The approach delivers real‑time performance (55.9 FPS on a single CPU thread) and outperforms lightweight CNN baselines by a large margin in both speed and efficiency.