Page tree
Skip to end of metadata
Go to start of metadata

Workshop Description and Motivation

Visual place recognition approaches are of fundamental importance to autonomous systems and vehicles such as personal service robots, self-driving cars, and unmanned aerial systems, as they allow these systems to navigate in the world. A major challenge for visual place recognition systems is to achieve robustness to the large variability in scene appearance that can be observed in the real world. Such changes – induced by the time of day, weather or seasonal effects as well as human activity – are a ubiquitous challenge for all autonomous systems aiming at long-term operations in both indoor and outdoor settings. Visual Place Recognition techniques need to be able to understand the scene and detect and react to changes in it, which requires concepts from image-based geo-localization, robotics, machine learning, and visual neuroscience.

The goal of this workshop is to bring together researchers working in different fields of computer vision, machine learning, robotics and visual neuroscience to discuss novel concepts and ideas for robust vision-based place recognition in severely changing environments.

The workshop begins with a brief tutorial that introduces the basic principles and the state of the art to participants without previous experience in the field. Besides oral and poster presentations, we have two invited talks by renowned experts in the fields of computer vision and mobile robotics. The workshop will also address questions of collecting long-term datasets in changing indoor and outdoor environments to enable standardized benchmarks.

Finally, we will run a place recognition challenge in association with the workshop, with prizes to be awarded at the conclusion of the workshop.

Anticipated Target Audience

The tutorial aims at both students with no experience in this research area and experts by providing a short tutorial on place recognition in changing environments and a place recognition challenge. The topics covered by this workshop touch many different research areas such as geometric computer vision, image-based localization and visual place recognition, robotics, machine learning, and visual neuroscience and should thus be interesting to a wide audience.

Important Dates

  • May 11: Extended deadline for submission of contributed papers
  • May 15: Notification of accepted papers (submitted papers will be reviewed timely and notification will be given as soon as possible to ease planning of conference travel for the participants)
  • Thursday, June 11: Workshop held at CVPR (morning session)

Call for Papers and Submission Guidelines

Papers can be submitted by e-mail to niko.suenderhauf@qut.edu.au and should follow the usual CVPR guidelines for style and length (up to 8 pages + references). The papers will be reviewed and commented by the members of the program committee. The accepted papers will be published on the workshop website. Papers will be presented in a short talk (expect 3-5 minutes) followed by a poster session.

Topics of Interest

Topics of interest to the workshop include, but are not limited to:

  • Visual localization and long-term mapping under severe appearance and viewpoint changes

  • Novel techniques for change invariant scene matching, based on whole images or features

  • Learning, modelling and predicting systematic / repeating appearance changes over time

  • Learning invariant properties of places / invariant image features

  • Semantic understanding of changes

  • Collecting long-term datasets in changing indoor and outdoor environments

  • Benchmarking long-term mapping and localization

We explicitly encourage the submission of papers describing work in progress, preliminary results or novel concepts.

 Contributed Papers

AuthorsTitlePaperPosterSlides

Yukyung Choi, Namil Kim, Kibaek Park, Soonmin Hwang, Jae Shin Yoon, In So Kweon 

All-Day Visual Place Recognition: Benchmark Dataset and Baseline

PDF  

Edward Johns and Guang-Zhong Yang 

RANSAC with Geometric Cliques for Image Retrieval and Place Recognition

PDFposterslides

Dmytro Mishkin, Michal Perdoch, Jiri Matas

Place Recognition with WxBS Retrieval 

PDFposter 
Peer Neubert, Stefan Schubert, Peter Protzel

Exploiting intra Database Similarities for Selection of Place Recognition Candidates in Changing Environments 

PDF  

Ivan Sikirić, Karla Brkić, Josip Krapac, Siniša Sěgvić

Robust Traffic Scene Recognition with a Limited Descriptor Length 

PDF

posterslides

Workshop Program

The workshop will be held in the morning session on June 11.

TimeProgram Item
8.30Welcome and Introduction
8.35Visual Place Recognition - Tutorial and Overview
9.00Invited Talk – Josef Sivic (INRIA)
9.30Invited Talk – John Leonard (MIT CSAIL)
10.00Coffee Break
10.30Short Presentations of Contributed Papers (5 minutes each)
10.30Choi et al. – All-Day Visual Place Recognition: Benchmark Dataset and Baseline
10.35Johns & Yang – RANSAC with Geometric Cliques for Image Retrieval and Place Recognition
10.40Mishkin et al. – Place Recognition with WxBS Retrieval 
10.45Neubert et al. – Exploiting intra Database Similarities for Selection of Place Recognition Candidates in Changing Environments 

10.50

Sikirić et al. – Robust Traffic Scene Recognition with a Limited Descriptor Length 
10.55Poster Session of Contributed Papers
11.50Open Discussion and Conclusions
12.00Wrap-up / Lunch

The CVPR Place Recognition Challenge

In conjunction with the workshop at CVPR we will organize a place recognition challenge. Participants can test their algorithms for visual place recognition on a challenging dataset and compare results with other participants.  

We encourage you to participate in this challenge and send in your results. Please follow this link for information about the challenge and the dataset.

Workshop Organizers

The workshop is organized by Dr Niko Sünderhauf, Dr Michael Milford, and Professor Peter Corke who are with the Australian Centre for Robotic Vision at QUT in Brisbane, Australia and Dr Torsten Sattler from ETH Zürich.

Contact details:

Dr Niko Sünderhauf
niko.suenderhauf@qut.edu.au
School of Electrical Engineering and Computer Science
Queensland University of Technology
Brisbane QLD 4000, Australia