MultiPoseNet: Fast Multi-Person Pose Estimation using Pose Residual Network

TL;DR

MultiPoseNet is a fast, multi-task bottom-up architecture for multi-person pose estimation that outperforms previous methods in accuracy and speed, achieving real-time performance at 23 frames/sec.

Contribution

Introduction of MultiPoseNet, a multi-task model with Pose Residual Network for joint detection and pose estimation, offering improved accuracy and speed over existing methods.

Findings

Outperforms previous bottom-up methods by +4-point mAP on COCO dataset

Achieves real-time performance at 23 frames/sec

Comparable accuracy to top-down methods while being at least 4x faster

Abstract

In this paper, we present MultiPoseNet, a novel bottom-up multi-person pose estimation architecture that combines a multi-task model with a novel assignment method. MultiPoseNet can jointly handle person detection, keypoint detection, person segmentation and pose estimation problems. The novel assignment method is implemented by the Pose Residual Network (PRN) which receives keypoint and person detections, and produces accurate poses by assigning keypoints to person instances. On the COCO keypoints dataset, our pose estimation method outperforms all previous bottom-up methods both in accuracy (+4-point mAP over previous best result) and speed; it also performs on par with the best top-down methods while being at least 4x faster. Our method is the fastest real time system with 23 frames/sec. Source code is available at: https://github.com/mkocabas/pose-residual-network