Human 3D keypoints via spatial uncertainty modeling

TL;DR

This paper presents a novel 3D human keypoint estimation method that models spatial uncertainty without requiring 3D ground truth, achieving near state-of-the-art results efficiently.

Contribution

It introduces a spatial uncertainty modeling approach for 3D keypoints that operates without 3D ground truth labels, inspired by robust statistics.

Findings

Achieves near state-of-the-art performance on Human3.6m

Does not require 3D ground truth labels

Efficient and straightforward to evaluate

Abstract

We introduce a technique for 3D human keypoint estimation that directly models the notion of spatial uncertainty of a keypoint. Our technique employs a principled approach to modelling spatial uncertainty inspired from techniques in robust statistics. Furthermore, our pipeline requires no 3D ground truth labels, relying instead on (possibly noisy) 2D image-level keypoints. Our method achieves near state-of-the-art performance on Human3.6m while being efficient to evaluate and straightforward to