Synthetic humans improve multi-person pose estimation mainly through extra occlusion; masking their losses beats using their labels, and a teacher-guided sampler adds a small extra gain.
In: International Conference on Learning Representations (ICLR)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Learning to Train with Synthetic Humans
Synthetic humans improve multi-person pose estimation mainly through extra occlusion; masking their losses beats using their labels, and a teacher-guided sampler adds a small extra gain.