Learning hough transform with latent structures for joint object detection and pose estimation

Hanxi Li*, Xuming He, Nick Barnes, Mingwen Wang

*Corresponding author for this work

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

1 Citation (Scopus)

Abstract

We present a novel max-margin Hough transform with latent structure for joint object detection and pose estimation. Our method addresses the large appearance and shape variation of objects in multiple poses by integrating three key components: First, we propose a more robust appearance model by designing a patch dictionary with complementary features; In addition, we use a group of latent components to explicitly incorporate feature selection and pooling into the Hough-based object models; Furthermore, we adopt a multiple instance learning approach to handle the lack of correspondence among training instances with noisy bounding-box labels. We design a unified objective and an efficient approximate inference that alternates the search between object location and pose space. We demonstrate the efficacy of our approach by achieving the state-of-the-art performance on two detection and two joint estimation datasets.

Original languageEnglish
Title of host publicationMultiMedia Modeling - 22nd International Conference, MMM 2016, Proceedings
EditorsRichang Hong, Nicu Sebe, Qi Tian, Guo-Jun Qi, Benoit Huet, Xueliang Liu
PublisherSpringer Verlag
Pages116-129
Number of pages14
ISBN (Print)9783319276731
DOIs
Publication statusPublished - 2016
Externally publishedYes
Event22nd International Conference on MultiMedia Modeling, MMM 2016 - Miami, United States
Duration: 4 Jan 20166 Jan 2016

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9517
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference22nd International Conference on MultiMedia Modeling, MMM 2016
Country/TerritoryUnited States
CityMiami
Period4/01/166/01/16

Fingerprint

Dive into the research topics of 'Learning hough transform with latent structures for joint object detection and pose estimation'. Together they form a unique fingerprint.

Cite this