Abstract

Convolutional neural networks (CNNs) have demonstrated remarkable success in vision-related tasks. However, their susceptibility to failing when inputs deviate from the training distribution is well-documented. Recent studies suggest that CNNs exhibit a bias toward texture instead of object shape in image classification tasks, and that background information may affect predictions. This paper investigates the ability of CNNs to adapt to different color distributions of an image while maintaining context and background. The results of our experiments on modified MNIST, CIFAR10 and CIFAR 100 data demonstrate that changes in color can substantially affect classification accuracy. The paper explores the effects of various regularization techniques on generalization error across datasets and proposes an architectural modification using in a novel way color balancing and spatial dropout regularization. This enhances the model reliance on color-invariant intensity-based features for improved classification accuracy. Overall, this work contributes to ongoing efforts to understand the limitations and challenges of CNNs in image classification tasks and offers a potential solution to improve their performance.

Statistics