Bidirectional Regression for Monocular 6DoF Head Pose Estimation and Reference System Alignment

TL;DR

This paper introduces TRGv2, a bidirectional regression network for monocular 6DoF head pose estimation that improves accuracy, generalization, and fairness in cross-dataset evaluation through iterative refinement and bias correction.

Contribution

We propose TRGv2, a novel bidirectional head pose estimation method with iterative refinement and bias correction for fair cross-dataset comparison.

Findings

TRGv2 outperforms state-of-the-art methods in accuracy.

TRGv2 demonstrates improved generalization to out-of-distribution data.

Bias correction enables fair evaluation across different datasets.

Abstract

Precise six-degree-of-freedom (6DoF) head pose estimation is crucial for safety-critical applications and human-computer interaction scenarios, yet existing monocular methods still struggle with robust pose estimation. We revisit this problem by introducing TRGv2, a lightweight extension of our previous Translation, Rotation, and Geometry (TRG) network, which explicitly models the bidirectional interaction between facial geometry and head pose. TRGv2 jointly infers facial landmarks and 6DoF pose through an iterative refinement loop with landmark-to-image projection, ensuring metric consistency among face size, rotation, and depth. To further improve generalization to out-of-distribution data, TRGv2 regresses correction parameters instead of directly predicting translation, combining them with a pinhole camera model for analytic depth estimation. In addition, we identify a previously overlooked source of bias in cross-dataset evaluations due to inconsistent head center definitions across different datasets. To address this, we propose a reference system alignment strategy that quantifies and corrects translation bias, enabling fair comparisons across datasets. Extensive experiments on ARKitFace, BIWI, and the challenging DD-Pose benchmarks demonstrate that TRGv2 outperforms state-of-the-art methods in both accuracy and efficiency. Code and newly annotated landmarks for DD-Pose will be publicly available.