Last Update : November 11, 2012

The 3rd International Workshop on

Benchmark Test Schemes for AR/MR

Geometric Registration and Tracking Method


Online Proceedings

Invited Talk
09:30 - 10:30Dense Localization and Mapping: Recent Advances and Benchmark Evaluation
Speaker: Daniel Cremers
Professor for Computer Science and Mathematics
Technische Universitat Munchen Abstract: Image-based reconstruction from color- or RGB-D cameras is an important challenge with numerous applications. I will present methods to tackle the problems of localization and mapping by means of spatially dense approaches. To this end, I will introduce variational approaches and show how they can be minimized using convex optimization techniques. In particular, I will focus on algorithms which allow realtime performance. Finally, I will discuss recent developments aiming at benchmarking the performance of such algorithms.

This is joint work with Kalin Kolev, Jan Stuehmer, Frank Steinbruecker and Juergen Sturm.

RGB-D benchmark:
Algorithm for realtime localization using RGB-D camera:
Algorithm for realtime dense reconstruction from a color camera:
Coffee Break
Tracking and evaluation
10:40 - 11:00Pose Estimation Evaluation of Viewpoint Generative Learning
Takumi Yoshida, Hideo Saito (Keio University)
11:00 - 11:20Real-time Line-based SLAM for AR
Keisuke Hirose, Hideo Saito (Keio University)
11:20 - 11:40Trial Evaluation of a Marker-based Geometric Registration Method Using a TrakMark Data Set
Yukiko Kubo, Shinya Komurasaki, Fumihisa Shibata, Asako Kimura, Hideyuki Tamura (Ritsumeikan University)
Pose estimation and visual servo
11:40 - 12:00Accurate Detection of Passive Marker for Pose Reconstruction in Dental Navigation Systems
Evgeniy Minaev, Artem Nikonorov (Russian Academy of Science)
12:00 - 12:20Application of Moiré Patterns to AR Markers for High-accuracy Pose Estimation
Hideyuki Tanaka, Yasushi Sumi, Yoshio Matsumoto (AIST)
12:20 - 12:40Development of Visual Servo Microscopes for Tracking of Model Organisms
Shogo Arai, Mitsunori Maru, Shintaro Matsui, Koichi Hashimoto (Tohoku University)
Lunch Time
Tracking of moving objects
13:40 - 14:00Visual Tracking Using L1 Minimization and Forward-Backward Verification
Dongjing Shan, Chao Zhang, Tan Xiao (Peking University)
14:00 - 14:20Occlusion Robust Recognition and Tracking of Motion Objects
Takashi Matsuzaki, Yuichi Yaguchi, Ryuichi Oka (University of Aizu)
14:20 - 14:40Tracking Sport Video Shots Region of Interest using the Angular Orientation Partition Descriptor
António M. G. Pinheiro (University of Beira Interior)
AR Benchmark
14:40 - 15:00A Benchmark Suite with Virtualized Reality Models for Supporting Tracking Evaluation and Data Set Generation
Koji Makita, Takashi Okuma, Tomoya Ishikawa (AIST), Laurence Nigay (UJF), Takeshi Kurata (AIST)
15:00 - 15:20Uniform Software Environment for AR Performance Evaluation based on USB boot Linux
Yoshinari Kameda, Itaru Kitahara, Yuichi Ohta (University of Tsukuba)
15:20 - 15:40Realistic CG Stereo Image Dataset with Ground Truth Disparity Maps
Sara Martull (University of Tsukuba), Martin Peris (Cyberdyne), Kazuhiro Fukui (University of Tsukuba)
Coffee Break
Panel Session
15:50 - 16:50Trakmark and beyond
This panel overviews the recent trend of camera tracking evaluation and discuss the future of highly and toughly evaluated camera tracking methods. In order to promote higher quality of camera tracking methodologies, what the TrakMark can do? What should be tackled first? Outstanding researchers in this literature will reveal some hints to answer these questions and have a hot dicussion.

Yoshinari Kameda (University of Tsukuba)

Daniel Cremers (Recent trend of RGB + D camera use)
Hideo Saito (Computer Vision and trend in ACCV)
Kiyoshi Kiyokawa (AR and trend in ISMAR)
Takeshi Kurata (Standardization)
16:50 - 17:00Hirokazu Kato (NAIST)