arXiv:2605.20624v1 Announce Type: cross Abstract: Diffusion models provide powerful priors for zero-shot video inverse problems, but their real-time deployment is hindered by two inefficiencies: high initial latency caused by holistic video restoration, and low throughput resulti…
Recent advances in Image Restoration (IR) have been largely driven by generative methods such as Diffusion Models and Flow Matching, which excel in synthesizing realistic textures while suffering from slow multi-step inference and compromised pixel fidelity. In contrast, classica…
Diffusion models provide powerful priors for zero-shot video inverse problems, but their real-time deployment is hindered by two inefficiencies: high initial latency caused by holistic video restoration, and low throughput resulting from multiple VAE passes to enforce measurement…
Diffusion-based posterior sampling (PS) is a leading framework for imaging inverse problems, combining learned priors with measurement constraints. Yet, its standard formulations rely on instantaneous data-consistent estimates, which induce temporal variability in the reverse dyn…
arXiv:2507.05482v3 Announce Type: replace-cross Abstract: Training-free diffusion guidance offers a flexible framework for leveraging off-the-shelf classifiers without additional training. Yet, current approaches hinge on posterior approximations via Tweedie's formula, which ofte…
arXiv cs.CV
TIER_1English(EN)·Nan Yang, Julian Straub, Fan Zhang, Richard Newcombe, Jakob Engel, Lingni Ma·
arXiv:2605.05390v1 Announce Type: new Abstract: Tracking 3D human motion from egocentric multi-camera headset is challenged by severe egomotion, partial visibility or occlusions and lack of training data. Existing methods designed for monocular video often require static or slowl…