NIRMAL Pooling: An Adaptive Max Pooling Approach with Non-linear Activation for Enhanced Image Classification

TL;DR

NIRMAL Pooling introduces an adaptive, non-linear pooling layer for CNNs that improves image classification accuracy across multiple datasets by dynamically adjusting pooling parameters and applying ReLU activation.

Contribution

The paper proposes NIRMAL Pooling, a novel adaptive pooling method with non-linear activation, enhancing CNN robustness and feature representation for image classification.

Findings

Achieves higher accuracy than standard max pooling on MNIST and CIFAR-10.

Demonstrates consistent performance improvements across diverse datasets.

Offers a flexible pooling alternative for CNN architectures.

Abstract

This paper presents NIRMAL Pooling, a novel pooling layer for Convolutional Neural Networks (CNNs) that integrates adaptive max pooling with non-linear activation function for image classification tasks. The acronym NIRMAL stands for Non-linear Activation, Intermediate Aggregation, Reduction, Maximum, Adaptive, and Localized. By dynamically adjusting pooling parameters based on desired output dimensions and applying a Rectified Linear Unit (ReLU) activation post-pooling, NIRMAL Pooling improves robustness and feature expressiveness. We evaluated its performance against standard Max Pooling on three benchmark datasets: MNIST Digits, MNIST Fashion, and CIFAR-10. NIRMAL Pooling achieves test accuracies of 99.25% (vs. 99.12% for Max Pooling) on MNIST Digits, 91.59% (vs. 91.44%) on MNIST Fashion, and 70.49% (vs. 68.87%) on CIFAR-10, demonstrating consistent improvements, particularly on complex datasets. This work highlights the potential of NIRMAL Pooling to enhance CNN performance in diverse image recognition tasks, offering a flexible and reliable alternative to traditional pooling methods.