The 2019 IEEE GRSS Data Fusion Contest, organized by the IADF TC in collaboration with the Johns Hopkins University (JHU) and the Intelligence Advanced Research Projects Activity (IARPA), was opened on January 11, 2019. The final phase was opened on March 7 and closed on March 22. We received over 1100 submissions!
We would like to thank all the participants for their submissions. Evaluation and ranking were conducted by the Organizers, who carefully checked all valid submissions. The winners are reported below along with their approaches and metrics over all the undisclosed ground truth samples.
Single-view Semantic 3D Challenge
1st Place
Author: Saket Kunwar
Affiliation: NestAI, Nepal
Team Name: nest
Approach: An ensemble of a few varied backbones employed in a U-Net architecture
Metric (mIoU-3): 0.5571
2nd Place
Authors: Zhuo Zheng, Yanfei Zhong, and Junjue Wang
Affiliation: Wuhan University, China
Team Name: RSIDEA-WHU (Dadajuan in Codalab)
Approach: A pyramid on pyramid network based on an encoder-dual decoder framework
Metric (mIoU-3): 0.534
Pairwise Semantic Stereo Challenge
1st Place
Authors: Hongyu Chen, Manhui Lin, Hongyan Zhang, Guangyi Yang, Guisong Xia, Xianwei Zheng, and Liangpei Zhang
Affiliation: Wuhan University, China
Team Name: BurningAllthing
Approach: A modified version of Pyramid Stereo Matching Network (PSMNet) and Disparity Fusion Segmentation Net (DFSN)
Metric (mIoU-3): 0.7775
2nd Place
Authors: Rongjun Qin, Xu Huang, Wei Liu, and Changlin Xiao
Affiliation: The Ohio State University, US
Team Name: qin.324
Approach: U-Net and Pyramid Stereo Matching Network (PSMNet)
Metric (mIoU-3): 0.7724
Multi-view Semantic Stereo Challenge
1st Place
Authors: Pablo d’Angelo, Daniele Cerra, Seyed Majid Azimi, Nina Merkle, Jiaojiao Tian, Stefan Auer, Miguel Pato, Raquel de los Reyes, Xiangyu Zhuo, Ksenia Bittner, Thomas Krauss, and Peter Reinartz
Affiliation: DLR, Germany
Team Name: Panoptes
Approach: Semi-global matching and an ensemble of CNN classifiers with ad hoc detectors
Metric (mIoU-3): 0.7461
2nd Place
Authors: Rongjun Qin, Xu Huang, Wei Liu, and Changlin Xiao
Affiliation: The Ohio State University, US
Team Name: qin.324
Approach: Semi-global matching and U-Net
Metric (mIoU-3): 0.7300
3D Point Cloud Classification Challenge
1st Place
Authors: Lian Yanchao, Feng Tuo, and Zhou Jinliu
Affiliation: Xidian University, China
Team Name: fengzicai
Approach: U-type Pointnet++ network (UPNet)
Metric (mIoU): 0.9455
2nd Place
Authors: Jia Meixia, Wu Zhaoyang, and Li Aijin
Affiliation: Xidian University, China
Team Name: aijinli0613
Approach: Attention-SIFT Net (AS net) based on Pointnet++ and PointSIFT
Metric (mIoU): 0.9454
Each team will present their methods in a special session at the forthcoming IGARSS conference in Yokohama, Japan. Their papers will be included in the IGARSS proceedings.