dc.citation.conferencePlace |
UK |
- |
dc.citation.conferencePlace |
Aberdeen |
- |
dc.citation.title |
British Machine Vision Conference |
- |
dc.contributor.author |
Jeong, Uyoung |
- |
dc.contributor.author |
Baek, Seungryul |
- |
dc.contributor.author |
Chang, Hyung Jin |
- |
dc.contributor.author |
Kim, Kwang In |
- |
dc.date.accessioned |
2024-01-31T18:06:15Z |
- |
dc.date.available |
2024-01-31T18:06:15Z |
- |
dc.date.created |
2023-09-26 |
- |
dc.date.issued |
2023-11-20 |
- |
dc.description.abstract |
Single-stage multi-person human pose estimation (MPPE) methods have shown great performance improvements, but existing methods fail to disentangle features by individual instances under crowded scenes. In this paper, we propose a bounding box-level instance representation learning called BoIR, which simultaneously solves instance detection, instance disentanglement, and instance-keypoint association problems. Our new instance embedding loss provides a learning signal on the entire area of the image with bounding box annotations, achieving globally consistent and disentangled instance representation. Our method exploits multi-task learning of bottom-up keypoint estimation, bounding box regression, and contrastive instance embedding learning, without additional computational cost during inference. BoIR is effective for crowded scenes, outperforming state-of-the-art on COCO val (0.8 AP), COCO test-dev (0.5 AP), CrowdPose (4.9 AP), and OCHuman (3.5 AP). Code will be available at https://github.com/uyoung-jeong/BoIR |
- |
dc.identifier.bibliographicCitation |
British Machine Vision Conference |
- |
dc.identifier.uri |
https://scholarworks.unist.ac.kr/handle/201301/74443 |
- |
dc.language |
영어 |
- |
dc.publisher |
British Machine Vision Association |
- |
dc.title |
BoIR: Box-Supervised Instance Representation for Multi Person Pose Estimation |
- |
dc.type |
Conference Paper |
- |
dc.date.conferenceDate |
2023-11-20 |
- |