Profile-Specific 3DMM Regression from a Single Lateral Face Image

2026-05-03

AI Summary

This study addresses the challenge of accurately reconstructing full 3D facial geometry from a single extreme profile image, which suffers from severe occlusion and sparse geometric cues, thereby limiting its clinical utility in applications such as orthodontic cephalometry. The work presents the first systematic investigation of 3D Morphable Model (3DMM) regression under this challenging viewpoint. It introduces ProfileSynth, a geometry-conditioned synthetic data generation strategy, and designs a profile-specific FLAME regression architecture augmented with visibility-aware mandibular contour regularization. The proposed method establishes the first practical baseline for profile-based 3D face reconstruction, achieving high-fidelity, clinically viable 3D facial geometry from a single RGB profile image and effectively filling a critical gap in the existing literature.

Abstract

Single-image 3D face reconstruction is a core problem in computer vision, with important clinical applications such as cephalometric landmark analysis in orthodontics. Traditionally, this analysis relies on lateral X-ray imaging; however, frequent X-ray exposure is impractical due to radiation concerns. While recent research has explored detecting landmarks from lateral RGB images as an alternative, existing methods typically rely on 2D features such as the eyes, mouth, ears, and boundary silhouettes, failing to fully exploit the underlying 3D facial geometry spanning the facial profile and jawline, which is essential for accurate diagnosis. Meanwhile, although 3D face reconstruction from frontal views has seen significant progress, most learning-based 3D morphable model (3DMM) regressors are developed and benchmarked on near-frontal images, where appearance cues are abundant. In extreme profile views (yaw $\approx 90^\circ$), much of the face is occluded, and the available signal is dominated by boundary cues, making accurate 3D reconstruction challenging. In this paper, we bridge this gap with geometry-conditioned synthetic data and a simple profile-specific FLAME regression baseline for single lateral images. We introduce ProfileSynth, a dataset created by sampling FLAME shape and pose parameters in extreme yaw ranges and generating photorealistic profile images using a diffusion model conditioned on depth and normal maps. We further study a profile-specific baseline with visibility-aware jawline regularization. Our framework provides a practical baseline for "profile $\times$ 3DMM" reconstruction and a promising foundation for more accurate, non-invasive cephalometric analysis from lateral RGB images.

Problem

• 3D face reconstruction
• lateral face image
• 3D morphable model
• cephalometric analysis
• profile view

Innovation

• profile-specific 3DMM
• single-image 3D face reconstruction
• geometry-conditioned synthesis
• visibility-aware regularization
• cephalometric analysis