SEAL-pose: Enhancing 3D Human Pose Estimation via a Learned Loss for Structural Consistency

TL;DR

SEAL-pose introduces a learnable, data-driven loss function that enhances 3D human pose estimation by capturing complex structural dependencies, outperforming traditional methods with manual priors.

Contribution

It proposes a novel learnable loss-net that learns structural dependencies directly from data, improving 3D human pose estimation without manual priors.

Findings

Reduces per-joint errors across benchmarks

Outperforms models with explicit structural constraints

Effective in cross-dataset and in-the-wild scenarios

Abstract

3D human pose estimation (HPE) is characterized by intricate local and global dependencies among joints. Conventional supervised losses are limited in capturing these correlations because they treat each joint independently. Previous studies have attempted to promote structural consistency through manually designed priors or rule-based constraints; however, these approaches typically require manual specification and are often non-differentiable, limiting their use as end-to-end training objectives. We propose SEAL-pose, a data-driven framework in which a learnable loss-net trains a pose-net by evaluating structural plausibility. Rather than relying on hand-crafted priors, our joint-graph-based design enables the loss-net to learn complex structural dependencies directly from data. Extensive experiments on three 3D HPE benchmarks with eight backbones show that SEAL-pose reduces per-joint errors and improves pose plausibility compared with the corresponding backbones across all settings. Beyond improving each backbone, SEAL-pose also outperforms models with explicit structural constraints, despite not enforcing any such constraints. Finally, we analyze the relationship between the loss-net and structural consistency, and evaluate SEAL-pose in cross-dataset and in-the-wild settings.