DiffSF: Diffusion Models for Scene Flow Estimation

arXiv:2403.05327v1 Announce Type: new
Abstract: Scene flow estimation is an essential ingredient for a variety of real-world applications, especially for autonomous agents, such as self-driving cars and robots. While recent scene flow estimation approaches achieve a reasonable accuracy, their applicability to real-world systems additionally benefits from a reliability measure. Aiming at improving accuracy while additionally providing an estimate for uncertainty, we propose DiffSF that combines transformer-based scene flow estimation with denoising diffusion models. In the diffusion process, the ground truth …

abstract accuracy agents applications arxiv autonomous autonomous agents benefits cars cs.cv diffusion diffusion models driving flow reliability robots self-driving systems type world