Pose Prior Learner: Unsupervised Categorical Prior Learning for Pose Estimation

2024-10-04

AI Summary

This work addresses the absence of generalizable pose priors in unsupervised category-level pose estimation by proposing the first fully unsupervised pose prior learning paradigm. Methodologically, it introduces a hierarchical memory network that stores and distills prototypical pose parts in a disentangled manner; joint optimization of template transformation and image reconstruction enables iterative prototype regression, substantially improving occlusion robustness. Key contributions include: (1) the first formal framework for unsupervised category-level pose prior learning; and (2) a hierarchical memory-driven mechanism for prototype disentanglement and dynamic regression. Extensive evaluation on human and animal pose datasets demonstrates consistent superiority over state-of-the-art unsupervised baselines—particularly under severe occlusion, where pose estimation accuracy improves markedly. The entire pipeline operates without any manual annotations or supervised pose priors.

Technology Category

• Computer Vision: Biometrics, Face, Gesture & Pose
• Intelligent Robots: State Estimation
• Machine Learning: Unsupervised & Self-Supervised Learning

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

A prior represents a set of beliefs or assumptions about a system, aiding inference and decision-making. In this paper, we introduce the challenge of unsupervised categorical prior learning in pose estimation, where AI models learn a general pose prior for an object category from images in a self-supervised manner. Although priors are effective in estimating pose, acquiring them can be difficult. We propose a novel method, named Pose Prior Learner (PPL), to learn a general pose prior for any object category. PPL uses a hierarchical memory to store compositional parts of prototypical poses, from which we distill a general pose prior. This prior improves pose estimation accuracy through template transformation and image reconstruction. PPL learns meaningful pose priors without any additional human annotations or interventions, outperforming competitive baselines on both human and animal pose estimation datasets. Notably, our experimental results reveal the effectiveness of PPL using learned prototypical poses for pose estimation on occluded images. Through iterative inference, PPL leverages the pose prior to refine estimated poses, regressing them to any prototypical poses stored in memory. Our code, model, and data will be publicly available.

Problem

• Unsupervised categorical prior learning
• Pose estimation from images
• Improving accuracy without human annotations

Innovation

• Unsupervised categorical prior learning
• Hierarchical memory for pose storage
• Self-supervised pose estimation enhancement