Human Pose Estimation

PifPaf: Composite Fields for Human Pose Estimation

Sven Kreiss, Lorenzo Bertoni, Alexandre Alahi

We propose a new bottom-up method for multi-person 2D human pose estimation that is particularly well suited for urban mobility such as self-driving cars and delivery robots. The new method, PifPaf, uses a Part Intensity Field (PIF) to localize body parts and a Part Association Field (PAF) to associate body parts with each other to form full human poses. Our method outperforms previous methods at low resolution and in crowded, cluttered and occluded scenes thanks to (i) our new composite field PAF encoding fine-grained information and (ii) the choice of Laplace loss for regressions which incorporates a notion of uncertainty. Our architecture is based on a fully convolutional, single-shot, box-free design. We perform on par with the existing state-of-the-art bottom-up method on the standard COCO keypoint task and produce state-of-the-art results on a modified COCO keypoint task for the transportation domain.

CVPR’19 PDF, ITS’21 PDF, GitHub

 

 

Reference

Composite Relationship Fields with Transformers for Scene Graph Generation

G. Adaimi; D. Mizrahi; A. Alahi 

2023. IEEE/CVF Winter Conference on Applications of Computer Vision (WACV 2023), Waikoloa, Hawaii, United States, January 3-7, 2023. p. 52-64. DOI : 10.1109/WACV56688.2023.00014.

OpenPifPaf: Composite Fields for Semantic Keypoint Detection and Spatio-Temporal Association

S. Kreiss; L. Bertoni; A. Alahi 

Ieee Transactions On Intelligent Transportation Systems. 2022. Vol. 23, num. 8, p. 13498-13511. DOI : 10.1109/TITS.2021.3124981.

PifPaf: Composite Fields for Human Pose Estimation

S. Kreiss; L. Bertoni; A. Alahi 

2019-06-01. IEEE conference on computer vision and pattern recognition (CVPR), Long Beach, CA, Jun 16-20, 2019. p. 11969-11978. DOI : 10.1109/CVPR.2019.01225.