# Monocular 3D Human Pose Estimation by Generation and Ordinal Ranking

## TL;DR

This paper introduces a novel CVAE-based approach for monocular 3D human pose estimation that generates diverse pose samples and uses ordinal ranking to improve accuracy, achieving near state-of-the-art results.

## Contribution

It proposes a generative model conditioned on 2D poses and a ranking strategy to effectively estimate 3D human poses from monocular images, even without paired annotations.

## Key findings

- Achieves near state-of-the-art results with OrdinalScore.
- Attains state-of-the-art results with Oracle supervision.
- Performs competitively without paired image-3D annotations.

## Abstract

Monocular 3D human-pose estimation from static images is a challenging problem, due to the curse of dimensionality and the ill-posed nature of lifting 2D-to-3D. In this paper, we propose a Deep Conditional Variational Autoencoder based model that synthesizes diverse anatomically plausible 3D-pose samples conditioned on the estimated 2D-pose. We show that CVAE-based 3D-pose sample set is consistent with the 2D-pose and helps tackling the inherent ambiguity in 2D-to-3D lifting. We propose two strategies for obtaining the final 3D pose- (a) depth-ordering/ordinal relations to score and weight-average the candidate 3D-poses, referred to as OrdinalScore, and (b) with supervision from an Oracle. We report close to state of-the-art results on two benchmark datasets using OrdinalScore, and state-of-the-art results using the Oracle. We also show that our pipeline yields competitive results without paired image-to-3D annotations. The training and evaluation code is available at https://github.com/ssfootball04/generative_pose.