Neural Pruning for 3D Scene Reconstruction: Efficient NeRF Acceleration

2025-04-01

AI Summary

Neural Radiance Fields (NeRFs) suffer from high computational cost and long training times, hindering deployment on resource-constrained edge devices. To address this, we propose a neural pruning-based framework for efficient 3D scene reconstruction. We systematically compare diverse pruning strategies and introduce a novel coreset-driven pruning framework that jointly optimizes uniform sampling, importance-based pruning, and subset selection—enabling structural model compression while preserving reconstruction fidelity. Experiments demonstrate a 50% reduction in model size, a 35% acceleration in training speed, and only a marginal PSNR degradation of 0.8 dB. Our approach significantly enhances the practicality and deployability of NeRFs in edge-computing scenarios, establishing a new paradigm for efficient neural radiance field modeling.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Computer Vision: 3D Computer Vision
• Data Mining & Knowledge Management: Data Compression

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Neural Radiance Fields (NeRF) have become a popular 3D reconstruction approach in recent years. While they produce high-quality results, they also demand lengthy training times, often spanning days. This paper studies neural pruning as a strategy to address these concerns. We compare pruning approaches, including uniform sampling, importance-based methods, and coreset-based techniques, to reduce the model size and speed up training. Our findings show that coreset-driven pruning can achieve a 50% reduction in model size and a 35% speedup in training, with only a slight decrease in accuracy. These results suggest that pruning can be an effective method for improving the efficiency of NeRF models in resource-limited settings.

Problem

• Reduce NeRF model size via neural pruning
• Accelerate NeRF training time efficiently
• Maintain accuracy while pruning NeRF models

Innovation

• Neural pruning for NeRF acceleration
• Coreset-driven pruning reduces model size
• Speeds up training with minimal accuracy loss