Multi-Domain Pose Network for Multi-Person Pose Estimation and Tracking

TL;DR

This paper introduces the Multi-Domain Pose Network (MDPN), a multi-domain learning approach that improves multi-person pose estimation and tracking by effectively utilizing multiple datasets, achieving top results without extra data.

Contribution

The paper proposes a novel multi-domain learning framework for pose estimation that enhances representation learning and improves performance on multi-person pose tracking tasks.

Findings

Achieves state-of-the-art performance on PoseTrack ECCV 2018 Challenge

Effectively leverages multiple datasets without additional data

Significantly outperforms baseline models

Abstract

Multi-person human pose estimation and tracking in the wild is important and challenging. For training a powerful model, large-scale training data are crucial. While there are several datasets for human pose estimation, the best practice for training on multi-dataset has not been investigated. In this paper, we present a simple network called Multi-Domain Pose Network (MDPN) to address this problem. By treating the task as multi-domain learning, our methods can learn a better representation for pose prediction. Together with prediction heads fine-tuning and multi-branch combination, it shows significant improvement over baselines and achieves the best performance on PoseTrack ECCV 2018 Challenge without additional datasets other than MPII and COCO.