PifPaf: Composite Fields for Human Pose Estimation
This addresses pose estimation for self-driving cars and delivery robots, offering incremental improvements in specific domains.
The paper tackles multi-person 2D human pose estimation in challenging urban mobility scenarios by proposing PifPaf, a bottom-up method using composite fields, which outperforms previous methods at low resolution and in crowded scenes and achieves state-of-the-art results on a modified COCO task for transportation.
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.