HyPlaneHead: Rethinking Tri-plane-like Representations in Full-Head Image Synthesis

2025-09-20

AI Summary

This work addresses three critical limitations of triplane representations in 3D-aware GANs for full-head portrait synthesis: (1) feature entanglement causing mirror artifacts; (2) inefficient feature utilization and poor detail generation due to non-uniform mapping from square feature grids to spherical geometry; and (3) inter-plane interference from cross-channel feature leakage. To this end, we propose Hybrid Plane (Hy-Plane), a novel representation. Its key contributions include: the first systematic analysis of triplane deficiencies; a hybrid coordinate system integrating planar locality with spherical global consistency; near-equal-area spherical parameterization to improve mapping uniformity; and a single-channel unified feature map to suppress inter-channel leakage. Integrated into standard 3D-aware GAN frameworks, Hy-Plane significantly enhances geometric consistency and texture fidelity, achieving state-of-the-art performance on full-head portrait synthesis—outperforming both conventional triplane and spherical triplane baselines across all metrics.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Tri-plane-like representations have been widely adopted in 3D-aware GANs for head image synthesis and other 3D object/scene modeling tasks due to their efficiency. However, querying features via Cartesian coordinate projection often leads to feature entanglement, which results in mirroring artifacts. A recent work, SphereHead, attempted to address this issue by introducing spherical tri-planes based on a spherical coordinate system. While it successfully mitigates feature entanglement, SphereHead suffers from uneven mapping between the square feature maps and the spherical planes, leading to inefficient feature map utilization during rendering and difficulties in generating fine image details. Moreover, both tri-plane and spherical tri-plane representations share a subtle yet persistent issue: feature penetration across convolutional channels can cause interference between planes, particularly when one plane dominates the others. These challenges collectively prevent tri-plane-based methods from reaching their full potential. In this paper, we systematically analyze these problems for the first time and propose innovative solutions to address them. Specifically, we introduce a novel hybrid-plane (hy-plane for short) representation that combines the strengths of both planar and spherical planes while avoiding their respective drawbacks. We further enhance the spherical plane by replacing the conventional theta-phi warping with a novel near-equal-area warping strategy, which maximizes the effective utilization of the square feature map. In addition, our generator synthesizes a single-channel unified feature map instead of multiple feature maps in separate channels, thereby effectively eliminating feature penetration. With a series of technical improvements, our hy-plane representation enables our method, HyPlaneHead, to achieve state-of-the-art performance in full-head image synthesis.

Problem

• Tri-plane representations cause feature entanglement and mirroring artifacts
• Spherical tri-planes suffer from uneven mapping and inefficient feature utilization
• Feature penetration across channels causes interference between different planes

Innovation

• Hybrid-plane representation combining planar and spherical planes
• Near-equal-area warping strategy for efficient feature utilization
• Single-channel unified feature map eliminating feature penetration