PCM-NeRF: Probabilistic Camera Modeling for Neural Radiance Fields under Pose Uncertainty

2026-04-20

AI Summary

This work addresses the sensitivity of conventional Neural Radiance Fields (NeRF) to inaccurate camera poses, which often leads to severe reconstruction artifacts. Building upon SG-NeRF, the authors propose a probabilistic camera modeling framework that represents each camera pose as a learnable uncertainty distribution. A confidence-driven regularization loss is introduced to adaptively modulate the strength of gradient updates, thereby mitigating the adverse influence of low-quality or outlier views. Notably, this approach is the first to incorporate learnable pose variances directly into NeRF optimization and achieves state-of-the-art performance on complex scenes with significant pose noise—without requiring foreground masks. Experimental results demonstrate clear improvements over existing methods, particularly in terms of Chamfer Distance and F-Score metrics.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Neural surface reconstruction methods typically treat camera poses as fixed values, assuming perfect accuracy from Structure-from-Motion (SfM) systems. This assumption breaks down with imperfect pose estimates, leading to distorted or incomplete reconstructions. We present PCM-NeRF, a probabilistic framework that augments neural surface reconstruction with per-camera learnable uncertainty, built on top of SG-NeRF. Rather than treating all cameras equally throughout optimization, we represent each pose as a distribution with a learnable mean and variance, initialized from SfM correspondence quality. An uncertainty regularization loss couples the learned variance to view confidence, and the resulting uncertainty directly modulates the effective pose learning rate: uncertain cameras receive damped gradient updates, preventing poorly initialized views from corrupting the reconstruction. This lightweight mechanism requires no changes to the rendering pipeline and adds negligible overhead. Experiments on challenging scenes with severe pose outliers demonstrate that PCM-NeRF consistently outperforms state-of-the-art methods in both Chamfer Distance and F-Score, particularly for geometrically complex structures, without requiring foreground masks.

Problem

• Neural Radiance Fields
• Camera Pose Uncertainty
• Surface Reconstruction
• Structure-from-Motion
• Pose Estimation

Innovation

• Probabilistic Camera Modeling
• Pose Uncertainty
• Neural Radiance Fields
• Uncertainty Regularization
• Gradient Damping