A Light-Weight Monocular Depth Estimation with Edge-Guided Occlusion Fading Reduction

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Self-supervised monocular depth estimation methods suffer occlusion fading, which is a result of a lack of supervision by the ground truth pixels. A recent work introduced a post-processing method to reduce occlusion fading; however, the results have a severe halo effect. This work proposes a novel edge-guided post-processing method that reduces occlusion fading for self-supervised monocular depth estimation. We also introduce Atrous Spatial Pyramid Pooling with Forward-Path (ASPPF) into the network to reduce computational costs and improve inference performance. The proposed ASPPF-based network is lighter, faster, and better than current depth estimation networks. Our light-weight network only needs 7.6 million parameters and can achieve up to 67 frames per second for 256×512 inputs using a single nVIDIA GTX1080 GPU. The proposed network also outperforms the current state-of-the-art methods on the KITTI benchmark. The ASPPF-based network and edge-guided post-processing produces better results, both quantitatively and qualitatively than the competitors.

Original languageEnglish (US)
Title of host publicationAdvances in Visual Computing - 15th International Symposium, ISVC 2020, Proceedings
EditorsGeorge Bebis, Zhaozheng Yin, Edward Kim, Jan Bender, Kartic Subr, Bum Chul Kwon, Jian Zhao, Denis Kalkofen, George Baciu
PublisherSpringer Science and Business Media Deutschland GmbH
Pages69-81
Number of pages13
ISBN (Print)9783030645588
DOIs
StatePublished - 2020
Event15th International Symposium on Visual Computing, ISVC 2020 - San Diego, United States
Duration: Oct 5 2020Oct 7 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12510 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference15th International Symposium on Visual Computing, ISVC 2020
CountryUnited States
CitySan Diego
Period10/5/2010/7/20

Keywords

  • Atrous Spatial Pyramid Pooling
  • Edge-Guided post-processing
  • Monocular depth estimation

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'A Light-Weight Monocular Depth Estimation with Edge-Guided Occlusion Fading Reduction'. Together they form a unique fingerprint.

Cite this