Akapulu Labs logo Akapulu Labs Research

Human-Preference Fine-Tuning for 3D-Aware Face Geometry

Today’s digest highlights a new way to shape 3D face geometry directly from human preference signals, without text prompts, mesh priors, or explicit surface supervision. The result is stronger facial structure while preserving the original 2D appearance.

Human-Preference Fine-Tuning for 3D-Aware Face Geometry

Side-by-side comparison of EG3D 3D-aware face GAN geometry before (left) and after (right) fine-tuning with human-preference supervision, highlighting improved mesh quality. From Sculpting NeRF Geometry.

Digital Humans & 3D Faces

Sculpting NeRF Geometry

Sculpting NeRF Geometry

Sculpting NeRF Geometry: Human-Preference Fine-Tuning of a 3D-Aware Face GAN

This paper fine-tunes a pretrained 3D-aware face GAN's geometry using a reward model trained on human preferences directly from the radiance field's density values. It uniquely improves 3D facial geometry without relying on text prompts, mesh priors, or explicit surface supervision, while preserving 2D appearance.