Solving Missing-Annotation Object Detection with Background Recalibration Loss

Zhang Han
Zhang Han
Chen Fangyi
Chen Fangyi
Hao Qiqi
Hao Qiqi
Zhu Chenchen
Zhu Chenchen

ICASSP, pp. 1888-1892, 2020.

Cited by: 0|Views50
EI

Abstract:

This paper focuses on a novel and challenging detection scenario: A majority of true objects/instances is unlabeled in the datasets, so these missing-labeled areas will be regarded as the background during training. Previous art on this problem has proposed to use soft sampling to re-weight the gradients of RoIs based on the overlaps wi...More

Code:

Data:

Your rating :
0

 

Tags
Comments