Deep Multi-Task Networks For Occluded Pedestrian Pose Estimation

TL;DR

This paper introduces a multi-task deep learning framework that enhances pedestrian pose estimation, especially for occluded pedestrians, by leveraging domain adaptation across different datasets and improving multiple related tasks.

Contribution

It proposes a novel multi-task framework with unsupervised domain adaptation to improve occluded pedestrian pose estimation across diverse datasets.

Findings

Achieved state-of-the-art performance in pose estimation, detection, and segmentation.

Effectively handles occlusions in pedestrian pose estimation.

Demonstrated improved generalization across datasets.

Abstract

Most of the existing works on pedestrian pose estimation do not consider estimating the pose of an occluded pedestrian, as the annotations of the occluded parts are not available in relevant automotive datasets. For example, CityPersons, a well-known dataset for pedestrian detection in automotive scenes does not provide pose annotations, whereas MS-COCO, a non-automotive dataset, contains human pose estimation. In this work, we propose a multi-task framework to extract pedestrian features through detection and instance segmentation tasks performed separately on these two distributions. Thereafter, an encoder learns pose specific features using an unsupervised instance-level domain adaptation method for the pedestrian instances from both distributions. The proposed framework has improved state-of-the-art performances of pose estimation, pedestrian detection, and instance segmentation.