Articulated pose estimation with flexible mixtures-of-parts

TLDR

The authors propose a novel part‑model representation for human pose estimation in static images. Their method uses flexible mixtures of part templates and a tree‑structured mixture model that captures contextual co‑occurrence and spatial relations, enabling efficient dynamic‑programming optimization. Experiments on standard pose benchmarks show the approach achieves state‑of‑the‑art accuracy, outperforming prior methods by 50% and running orders of magnitude faster.

Abstract

We describe a method for human pose estimation in static images based on a novel representation of part models. Notably, we do not use articulated limb parts, but rather capture orientation with a mixture of templates for each part. We describe a general, flexible mixture model for capturing contextual co-occurrence relations between parts, augmenting standard spring models that encode spatial relations. We show that such relations can capture notions of local rigidity. When co-occurrence and spatial relations are tree-structured, our model can be efficiently optimized with dynamic programming. We present experimental results on standard benchmarks for pose estimation that indicate our approach is the state-of-the-art system for pose estimation, outperforming past work by 50% while being orders of magnitude faster.