# PifPaf: Composite Fields for Human Pose Estimation

## TL;DR

PifPaf introduces a novel bottom-up approach for multi-person 2D human pose estimation, excelling in crowded and occluded scenes by using composite fields and uncertainty-aware regression, suitable for urban mobility applications.

## Contribution

It presents PifPaf, a new fully convolutional, single-shot method with composite Part Intensity and Association Fields, improving pose estimation in challenging environments.

## Key findings

- Outperforms previous methods at low resolution and in crowded scenes.
- Achieves state-of-the-art results on transportation domain datasets.
- Performs on par with top methods on standard COCO keypoint task.

## Abstract

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.