Tianfei Zhou [1], Wenguan Wang[1], Zhijie Zhang [2], Jianbing Shen [1]. [1] Inception Institute of Artificial Intelligence (IIAI) [2] Tianjin University
We propose a Unified Relationship detection Network (URNet) in this contest. We jointly train entity detection/segmentation and relationship inference in an end-to-end way. Our URNet is built on Cascade (Mask)RCNN, which is further extended to add a new branch for relationship recognition for each human-object pair. The relationship recognition branch operates on rich multi-modal features for accurate inference.
Oct. 17, 2019, 6:30 p.m.
HTC+iCAN
0.5621
0.5232
0.3749
0.4867
Contributors
Description
Wenbo Song [1] [1] Tianjin University
We combined Hybrid Task Cascade with iCAN to generate the scene graph.
Oct. 16, 2019, 5:22 p.m.
RelNet
0.5317
0.4926
0.3244
0.4496
Contributors
Description
Kai Chen[1], Chao Xu[1], Jiahui Li[1] [1] Sun Yat-sen University
