English

Pixel-Pair Occlusion Relationship Map(P2ORM): Formulation, Inference & Application

Computer Vision and Pattern Recognition 2020-07-24 v1

Abstract

We formalize concepts around geometric occlusion in 2D images (i.e., ignoring semantics), and propose a novel unified formulation of both occlusion boundaries and occlusion orientations via a pixel-pair occlusion relation. The former provides a way to generate large-scale accurate occlusion datasets while, based on the latter, we propose a novel method for task-independent pixel-level occlusion relationship estimation from single images. Experiments on a variety of datasets demonstrate that our method outperforms existing ones on this task. To further illustrate the value of our formulation, we also propose a new depth map refinement method that consistently improve the performance of state-of-the-art monocular depth estimation methods. Our code and data are available at http://imagine.enpc.fr/~qiux/P2ORM/.

Keywords

Cite

@article{arxiv.2007.12088,
  title  = {Pixel-Pair Occlusion Relationship Map(P2ORM): Formulation, Inference & Application},
  author = {Xuchong Qiu and Yang Xiao and Chaohui Wang and Renaud Marlet},
  journal= {arXiv preprint arXiv:2007.12088},
  year   = {2020}
}

Comments

Accepted to ECCV 2020 as a spotlight. Project page: http://imagine.enpc.fr/~qiux/P2ORM/

R2 v1 2026-06-23T17:21:10.866Z