Multi-view pose estimation with mixtures-of-parts and adaptive viewpoint selection

TL;DR

This paper introduces a multi-view human pose estimation method that combines geometric and appearance coherence, adaptively weighting views to improve accuracy over single-view approaches.

Contribution

It presents a novel multi-view pose estimation framework that models view coherence and adaptively adjusts view contributions using latent appearance variables.

Findings

Significantly reduces pose estimation error compared to single-view methods

Effectively models geometric and appearance coherence across multiple views

Demonstrates robustness on HumanEva and UMPM datasets

Abstract

We propose a new method for human pose estimation which leverages information from multiple views to impose a strong prior on articulated pose. The novelty of the method concerns the types of coherence modelled. Consistency is maximised over the different views through different terms modelling classical geometric information (coherence of the resulting poses) as well as appearance information which is modelled as latent variables in the global energy function. Moreover, adequacy of each view is assessed and their contributions are adjusted accordingly. Experiments on the HumanEva and UMPM datasets show that the proposed method significantly decreases the estimation error compared to single-view results.