Using Data Augmentation to Improve the Robustness of Image Classification Models to Common Perturbations

Author(s)	Kapeesh Gudla
Country	India
Abstract	This study identified the improvement of robust accuracy of image classification models by using image perturbations. There are 7 perturbations which are added to images in the CIFAR-10 dataset, after which the model is subsequently tested for its accuracy in identifying images in test data. The research agrees that using perturbations can improve the robustness of image classification models. This paper emphasises on understanding and evaluating how image perturbations which are added to the training dataset can allow Image Classifiers to fare far better on the test data set. In terms of the bigger picture, this method of training a model can lead to fewer discriminatory outcomes when employed in the real-world. This is because concerns, such as racism and sexism, and other concerns regarding the functioning of Image Classifiers are reduced. Additionally, self-driving cars could become more efficient, and CCTVs equipped with these models can easily track different objects in a particular scene.
Keywords	Robustness; Data Augmentation; Perturbations; Image Classification; Data Bias; Adversarial Examples; Convolutional Neural Networks
Field	Computer > Artificial Intelligence / Simulation / Virtual Reality
Published In	Volume 8, Issue 1, January-February 2026
Published On	2026-01-17
DOI	https://doi.org/10.36948/ijfmr.2026.v08i01.66667

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research