DeblurNVS: Geometric Latent Diffusion for Novel View Synthesis from Sparse Motion-Blurred Images
Researchers have developed DeblurNVS, a new framework designed to synthesize high-fidelity novel views from sparse, motion-blurred images without needing per-scene optimization. This method works by restoring intermediate geometric representations that help recover structure and correspondence cues from the blurred inputs. DeblurNVS has demonstrated superior performance on synthetic motion-blur benchmarks and shows effectiveness on real-world blurred scenes, producing sharper and more stable novel views. AI
IMPACT Enables high-quality novel view synthesis from degraded imagery, potentially improving applications in robotics and augmented reality.