PASCAL in Detail
Make PASCAL great again!
CVPR 2017 Workshop Challenge
PASCAL in Detail Workshop Challenge, July 26th 2017 @ Room 323C, Hawaii Convention Center
DATASET ROLLOUT (AS OF 7/26)
- Image Classification [RELEASED]
- Object Detection [RELEASED]
- Semantic Segmentation [RELEASED]
- Instance Segmentation [RELEASED]
- Object Part Segmentation [RELEASED]
- Objectness Estimation [RELEASED]
- Boundary Detection [RELEASED]
- Occlusion Recognition [RELEASED]
- Human Keypoint Estimation [RELEASED]
- Human Action Recognition
The latest published dataset, trainval_withkeypoints.json, was published on July 26.
Check out the Dataset page for the latest news!
Overview
The goal of this workshop-challenge is to measure the progress in image understanding as reflected in a diverse set of visual tasks, capturing both low- and high-level aspects of vision problems. To accomplish this we systematize the evaluation of a large selection of representative visual tasks using carefully collected manual annotations. We thereby aspire to promote research that pushes the performance envelope in all facets of current computer vision research.
We will be having separate benchmarks for each of our competition tracks and winning entries will be invited to present their works.
We also introduce for the first time task triathlons for selected task combinations, as well as a task decathlon, where a single model will need to solve all 10 tasks combined. See the Multi-Task Challenges page for more info.
We also introduce taster challenges, including the Visual Domain Decathlon, where a single model will perform classification on ten different domains - lifting our eyes from flowers to airplanes.
Keynote speakers
Antonio Torralba (Professor at MIT)
Kevin Murphy (Research Scientist at Google)
Larry Zitnick (Research Manager at FAIR)
Kaiming He (Research Scientist at FAIR)
Schedule
(Updated: 07/26)
July 26th
9:00 - 9:30 Introduction and welcome (A. Yuille)
9:30 - 10:30 Single-track challenges and benchmarks
10:30 - 11:00 Invited keynote I (A. Torralba)
11:00 - 12:00 Invited keynote II (K. Murphy)
Lunch Break
14:00 - 14:45 Invited keynote III (L. Zitnick)
14:45 - 15:30 Decathlon challenges
15:30 - 16:00 Invited keynote IV (K. He)
Mask R-CNN
Single-Track Challenges
- Image Classification
- Object Detection
- Semantic Segmentation
- Instance Segmentation
- Object Part Segmentation
- Objectness
- Boundary Detection
- Occlusion Recognition
- Human Keypoint Estimation
- Human Action Recognition
Tasters
- Saliency Estimation
- Line Segment Detection
- Symmetry Detection
Multi-Task Challenges
- Boxes to Points Triathlon:
Object Detection, Instance Segmentation, Keypoint Estimation
- PASCAL++ Triathlon:
Image Classification, Object Detection, Semantic Segmentation
- Humans in Detail Triathlon:
Human Parts, Keypoints, Action
- PASCAL Decathlon:
All 10 tasks
Multi-Domain Challenge
VGG Domain Decathlon:
Image Classification in 10 Domains
Organizers and Technical Staffs
In alphabetical order of last names
- Sanja Fidler (Univ. Toronto)
- Iasonas Kokkinos (UCL; FAIR)
- Roozbeh Mottaghi (Allen Institute)
- George Papandreou (Google)
- Raquel Urtasun (Univ. Toronto)
- Andrea Vedaldi (Univ. Oxford)
- Alan L. Yuille (Johns Hopkins Univ.)
- Xianjie Chen (Facebook Research)
- Nam-Gyu Cho (Korean Univ.)
- Xiaodi Hou (TuSimple)
- Seong-Whan Lee (Korean Univ.)
- Yin Li (Gerogia Tech. Univ.)
- Xiaochen Lian (Baidu Research)
- Qing Liu (Johns Hopkins Univ.)
- Xiaobai Liu (San Diego State Univ.)
- Viet T. Nguyen (Jet Propulsion Lab)
- Vittal Premachandran (Johns Hopkins Univ.)
- Matthew Richard (Johns Hopkins Univ.)
- Wei Shen (Shanghai Univ.; Johns Hopkins Univ.)
- Peng Wang (Baidu Research)
- Xiang Xiang (Johns Hopkins Univ.)
- Zhuotun Zhu (Johns Hopkins Univ.)
Please do not hesitate to contact pascalindetail@gmail.com regarding any issues about this challenge.