SLIDE 1
Hang Zhao
Massachusetts Institute of Technology
Hang Zhao Massachusetts Institute of Technology Last years - - PowerPoint PPT Presentation
Hang Zhao Massachusetts Institute of Technology Last years - - PowerPoint PPT Presentation
Hang Zhao Massachusetts Institute of Technology Last years challenge at ECCV ADE Dataset New tasks this year Instance Segmentation Instance Segmentation 2.5 10 4 # instances Sta$s$cs: 2 100 foreground categories 200k
SLIDE 2
SLIDE 3
ADE Dataset
SLIDE 4
New tasks this year…
SLIDE 5
Instance Segmentation
SLIDE 6
10 20 30 40 50 60 70 80 90 100 0.5 1 1.5 2 2.5 ×104
Instance Segmentation
- Sta$s$cs:
- 100 foreground categories
- 200k instances
- #Instances per category: 200 ~ 20k
- #Instances per image: ~11
- Evalua$on:
- Mean AP over 100 categories
# instances
SLIDE 7
Semantic Boundary Detection
Perceptual Edges
SLIDE 8
Semantic Boundary Detection
Binary Edges Category-aware Semantic Boundaries
SLIDE 9
Semantic Boundary Detection
Boundary between: person and sofa Boundary between: person and chair Boundaries reveal object interac8ons
SLIDE 10
Semantic Boundary Detection
- 150 categories
- Inter-class boundaries
- Intra-class boundaries
- Evalua$on:
- F-measure at op$mal dataset scale (F-ODS)
SLIDE 11
Results – Scene Parsing
Team Name Final Score CASIA_IVA_JD 0.5547 WinterIsComing 0.5544 xdliang 0.5447 lijc15 0.5294 … … baseline-DilatedNet 0.4567
CASIA_IVA_JD Jun Fu, Jing Liu, Longteng Guo, Haijie Tian, Fei Liu, Yong Li, Yongjun Bao, Weipeng Yan, Hanqing Lu CASIA IVA and JD.com WinterIsComing Riwei Chen, Qi Chen, Xinglong Wu, Yifan Lu, Yudong Jiang, Linfu Wen Tou8ao AI Lab, ByteDance
Final Score = (mean IoU + pixel accuracy) / 2
* We received 30 submissions from 10 teams in this task. Full ranking can be found on places challenge website.
SLIDE 12
Sample Results Scene Parsing
SLIDE 13
Image Ground Truth Predic$on
SLIDE 14
Image Ground Truth Predic$on
SLIDE 15
Results – Instance Segmentation
Team Name mean AP Megvii (Face++) 0.2977 G-RMI 0.2415 BlueSky 0.1551 … … baseline 0.200
Megvii (Face++)
*Tete Xiao, *Ruixuan Luo, *Borui Jiang,
Shuai Shao, Yuning Jiang, Yadong Mu, Jieqi Shi, Chi Zhang, Jian Sun Megvii Research and Peking University G-RMI Alireza Fathi, Nori Kanazawa, Kai Yang, Kevin Murphy Google Research Machine Intelligence * We received 15 submissions from 5 teams in this task. Full ranking can be found on places challenge website.
SLIDE 16
Sample Results Instance Segmentation
SLIDE 17
SLIDE 18
SLIDE 19
Results – Semantic Boundary Detection
State-of-the-art: CASENet CVPR’17 hmp://www.merl.com/research/license#CASENet
- Z. Yu, et al. CASENet: Deep category-aware semantic edge detection
SLIDE 20
Sample Results Semantic Boundary Detection
SLIDE 21
Image Ground Truth Predic$on
SLIDE 22
Image Ground Truth Predic$on
SLIDE 23
Winner of Scene Parsing: CASIA_IVA_JD Jun Fu, Jing Liu, Longteng Guo, Haijie Tian, Fei Liu, Yong Li, Yongjun Bao, Weipeng Yan, Hanqing Lu CASIA IVA and JD.com
Awards
Prize: Titan Xp GPU Sponsored by: NVIDIA
SLIDE 24