Large Area 3D Human Pose Detection Via Stereo Reconstruction in Panoramic Cameras

TL;DR

This paper introduces a 3D human pose detection method using stereo panoramic cameras that transforms fisheye images into rectilinear views, enabling accurate pose estimation over large areas without retraining.

Contribution

The novel approach leverages panoramic cameras and image transformation to apply existing 2D pose estimation methods directly, simplifying 3D pose detection over wide fields of view.

Findings

Accurately estimates human poses over large areas.

Eliminates need for retraining deep learning models for fisheye distortions.

Suitable for ergonomic and pose-based assessments.

Abstract

We propose a novel 3D human pose detector using two panoramic cameras. We show that transforming fisheye perspectives to rectilinear views allows a direct application of two-dimensional deep-learning pose estimation methods, without the explicit need for a costly re-training step to compensate for fisheye image distortions. By utilizing panoramic cameras, our method is capable of accurately estimating human poses over a large field of view. This renders our method suitable for ergonomic analyses and other pose based assessments.