Fast3Dcache: Training-free 3D Geometry Synthesis Acceleration

2025-11-27

AI Summary

To address the slow inference of 3D diffusion models and geometric distortions or topological inconsistencies introduced by conventional caching strategies, this paper proposes a training-free, geometry-aware caching framework. Our method introduces predictive caching scheduling constraints and spatiotemporal stability criteria, integrating voxel-level stable pattern identification with velocity- and acceleration-driven feature stability assessment to dynamically allocate cache quotas and select reusable features. To our knowledge, this is the first approach that achieves effective acceleration while strictly preserving geometric consistency. Experiments demonstrate a 27.12% inference speedup, a 54.8% reduction in FLOPs, and only marginal degradation in reconstruction fidelity—specifically, a 2.48% increase in Chamfer Distance and a 1.95% drop in F-Score—thereby striking a compelling balance between efficiency and geometric fidelity.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Diffusion models have achieved impressive generative quality across modalities like 2D images, videos, and 3D shapes, but their inference remains computationally expensive due to the iterative denoising process. While recent caching-based methods effectively reuse redundant computations to speed up 2D and video generation, directly applying these techniques to 3D diffusion models can severely disrupt geometric consistency. In 3D synthesis, even minor numerical errors in cached latent features accumulate, causing structural artifacts and topological inconsistencies. To overcome this limitation, we propose Fast3Dcache, a training-free geometry-aware caching framework that accelerates 3D diffusion inference while preserving geometric fidelity. Our method introduces a Predictive Caching Scheduler Constraint (PCSC) to dynamically determine cache quotas according to voxel stabilization patterns and a Spatiotemporal Stability Criterion (SSC) to select stable features for reuse based on velocity magnitude and acceleration criterion. Comprehensive experiments show that Fast3Dcache accelerates inference significantly, achieving up to a 27.12% speed-up and a 54.8% reduction in FLOPs, with minimal degradation in geometric quality as measured by Chamfer Distance (2.48%) and F-Score (1.95%).

Problem

• Accelerates 3D diffusion inference while preserving geometric fidelity
• Reduces computational cost of iterative denoising in 3D synthesis
• Prevents structural artifacts from cached feature errors in 3D generation

Innovation

• Training-free caching framework accelerates 3D diffusion inference
• Predictive scheduler dynamically allocates cache based on voxel stability
• Spatiotemporal criterion selects stable features using velocity and acceleration