Learning from Noise: Enhancing DNNs for Event-Based Vision through Controlled Noise Injection

TL;DR

This paper introduces a noise-injection training method for event-based vision deep neural networks, improving robustness and accuracy in noisy conditions by directly incorporating controlled noise during training.

Contribution

The paper presents a novel noise-injection training approach that enhances deep learning models' robustness to event noise in event-based vision systems.

Findings

Outperforms traditional event-filtering techniques

Achieves stable performance across various noise levels

Provides highest average classification accuracy on benchmark datasets

Abstract

Event-based sensors offer significant advantages over traditional frame-based cameras, especially in scenarios involving rapid motion or challenging lighting conditions. However, event data frequently suffers from considerable noise, negatively impacting the performance and robustness of deep learning models. Traditionally, this problem has been addressed by applying filtering algorithms to the event stream, but this may also remove some of relevant data. In this paper, we propose a novel noise-injection training methodology designed to enhance the neural networks robustness against varying levels of event noise. Our approach introduces controlled noise directly into the training data, enabling models to learn noise-resilient representations. We have conducted extensive evaluations of the proposed method using multiple benchmark datasets (N-Caltech101, N-Cars, and Mini N-ImageNet) and various network architectures, including Convolutional Neural Networks, Vision Transformers, Spiking Neural Networks, and Graph Convolutional Networks. Experimental results show that our noise-injection training strategy achieves stable performance over a range of noise intensities, consistently outperforms event-filtering techniques, and achieves the highest average classification accuracy, making it a viable alternative to traditional event-data filtering methods in an object classification system. Code: https://github.com/vision-agh/DVS_Filtering