Skip to main navigation Skip to search Skip to main content

DOOBNet: Deep Object Occlusion Boundary Detection from an Image

  • Guoxia Wang
  • , Xiaochuan Wang
  • , Frederick W.B. Li
  • , Xiaohui Liang*
  • *Corresponding author for this work
  • Beihang University
  • Durham University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Object occlusion boundary detection is a fundamental and crucial research problem in computer vision. Solving this problem is challenging as we encounter extreme boundary/non-boundary class imbalance during the training of an object occlusion boundary detector. In this paper, we propose to address this class imbalance by up-weighting the loss contribution of false negative and false positive examples with our novel Attention Loss function. We also propose a unified end-to-end multi-task deep object occlusion boundary detection network (DOOBNet) by sharing convolutional features to simultaneously predict object boundary and occlusion orientation. DOOBNet adopts an encoder-decoder structure with skip connection in order to automatically learn multi-scale and multi-level features. We significantly surpass the state-of-the-art on the PIOD dataset (ODS F-score of.702) and the BSDS ownership dataset (ODS F-score of.555), as well as improving the detecting speed to as 0.037 s per image on the PIOD dataset.

Original languageEnglish
Title of host publicationComputer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers
EditorsGreg Mori, Hongdong Li, C.V. Jawahar, Konrad Schindler
PublisherSpringer Verlag
Pages686-702
Number of pages17
ISBN (Print)9783030208752
DOIs
StatePublished - 2019
Event14th Asian Conference on Computer Vision, ACCV 2018 - Perth, Australia
Duration: 2 Dec 20186 Dec 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11366 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference14th Asian Conference on Computer Vision, ACCV 2018
Country/TerritoryAustralia
CityPerth
Period2/12/186/12/18

Keywords

  • Boundary detection
  • Convolutional neural network
  • Occlusion reasoning

Fingerprint

Dive into the research topics of 'DOOBNet: Deep Object Occlusion Boundary Detection from an Image'. Together they form a unique fingerprint.

Cite this