A Simple And Effective Filtering Scheme For Improving Neural Fields

TL;DR

This paper introduces a novel filtering scheme for neural fields that combines smoothing and recovering operators to reduce noise while preserving details, improving generalization, controllability, and convergence.

Contribution

A new filtering technique that balances smoothing and sharpening operators to enhance local control and detail preservation in neural fields.

Findings

Significant noise reduction in neural fields.

Enhanced detail preservation and controllability.

Improved convergence speed and network stability.

Abstract

Recently, neural fields, also known as coordinate-based MLPs, have achieved impressive results in representing low-dimensional data. Unlike CNN, MLPs are globally connected and lack local control; adjusting a local region leads to global changes. Therefore, improving local neural fields usually leads to a dilemma: filtering out local artifacts can simultaneously smooth away desired details. Our solution is a new filtering technique that consists of two counteractive operators: a smoothing operator that provides global smoothing for better generalization, and conversely a recovering operator that provides better controllability for local adjustments. We have found that using either operator alone can lead to an increase in noisy artifacts or oversmoothed regions. By combining the two operators, smoothing and sharpening can be adjusted to first smooth the entire region and then recover fine-grained details in regions overly smoothed. In this way, our filter helps neural fields remove much noise while enhancing details. We demonstrate the benefits of our filter on various tasks and show significant improvements over state-of-the-art methods. Moreover, our filter also provides better performance in terms of convergence speed and network stability.