The Role of Noisy Data in Improving CNN Robustness for Image Classification

TL;DR

Introducing controlled noise into training data can significantly enhance CNN robustness for image classification, with minimal performance loss on clean data, by acting as an effective regularizer.

Contribution

This study demonstrates that adding a small proportion of noisy data during training improves CNN robustness against corruptions, a novel approach to regularization in image classification.

Findings

10% noisy data improves robustness significantly

Minimal impact on clean-data accuracy

Noise acts as an effective regularizer

Abstract

Data quality plays a central role in the performance and robustness of convolutional neural networks (CNNs) for image classification. While high-quality data is often preferred for training, real-world inputs are frequently affected by noise and other distortions. This paper investigates the effect of deliberately introducing controlled noise into the training data to improve model robustness. Using the CIFAR-10 dataset, we evaluate the impact of three common corruptions, namely Gaussian noise, Salt-and-Pepper noise, and Gaussian blur at varying intensities and training set pollution levels. Experiments using a Resnet-18 model reveal that incorporating just 10\% noisy data during training is sufficient to significantly reduce test loss and enhance accuracy under fully corrupted test conditions, with minimal impact on clean-data performance. These findings suggest that strategic exposure to noise can act as a simple yet effective regularizer, offering a practical trade-off between traditional data cleanliness and real-world resilience.