Inverse Rendering for High-Genus Surface Meshes from Multi-View Images

2025-11-23

AI Summary

Existing inverse rendering methods suffer from topological feature loss in high-genus surface reconstruction and geometric detail degradation in low-genus surfaces due to gradient instability induced by the Adam optimizer. To address this, we propose a topology-aware joint geometry–topology reconstruction framework. Leveraging the Gauss–Bonnet theorem, we formulate topological primitives that explicitly encode genus constraints. We further design an adaptive V-cycle remeshing scheme coupled with a reparameterized Adam optimizer to ensure gradient stability and multi-scale geometric fidelity. Our method integrates differential geometric modeling, dynamic mesh coarsening/refinement, and topology-aware regularization within the inverse rendering paradigm. Experiments demonstrate significant improvements over state-of-the-art methods in Chamfer Distance and Volume IoU. Crucially, our approach preserves topological consistency for high-genus surfaces while recovering fine-grained geometric details on low-genus surfaces.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Search and Optimization: Sampling/Simulation-based Search
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

We present a topology-informed inverse rendering approach for reconstructing high-genus surface meshes from multi-view images. Compared to 3D representations like voxels and point clouds, mesh-based representations are preferred as they enable the application of differential geometry theory and are optimized for modern graphics pipelines. However, existing inverse rendering methods often fail catastrophically on high-genus surfaces, leading to the loss of key topological features, and tend to oversmooth low-genus surfaces, resulting in the loss of surface details. This failure stems from their overreliance on Adam-based optimizers, which can lead to vanishing and exploding gradients. To overcome these challenges, we introduce an adaptive V-cycle remeshing scheme in conjunction with a re-parametrized Adam optimizer to enhance topological and geometric awareness. By periodically coarsening and refining the deforming mesh, our method informs mesh vertices of their current topology and geometry before optimization, mitigating gradient issues while preserving essential topological features. Additionally, we enforce topological consistency by constructing topological primitives with genus numbers that match those of ground truth using Gauss-Bonnet theorem. Experimental results demonstrate that our inverse rendering approach outperforms the current state-of-the-art method, achieving significant improvements in Chamfer Distance and Volume IoU, particularly for high-genus surfaces, while also enhancing surface details for low-genus surfaces.

Problem

• Reconstructs high-genus surface meshes from multi-view images
• Addresses gradient vanishing and exploding in mesh optimization
• Preserves topological features and surface details during reconstruction

Innovation

• Adaptive V-cycle remeshing scheme enhances topology awareness
• Re-parametrized Adam optimizer mitigates gradient vanishing issues
• Gauss-Bonnet theorem enforces topological consistency with genus numbers