Open Access

ARTICLE

Fuzzy C-Means Clustering-Driven Pooling for Robust and Generalizable Convolutional Neural Networks

Seunggyu Byeon¹, Jung-hun Lee², Jong-Deok Kim^3,*

1 Department of Computer Engineering, Dong-eui University, Busan, 47340, Republic of Korea
2 Division of Electrical and Electronic Engineering, Korea Maritime & Ocean University, Busan, 49112, Republic of Korea
3 Department of Information Convergence Engineering, Pusan National University, Busan, 46241, Republic of Korea

* Corresponding Author: Jong-Deok Kim. Email:

(This article belongs to the Special Issue: Recent Fuzzy Techniques in Image Processing and its Applications)

Computers, Materials & Continua 2026, 87(2), 24 https://doi.org/10.32604/cmc.2025.074033

Received 30 September 2025; Accepted 27 November 2025; Issue published 12 March 2026

Abstract

This paper introduces a fuzzy C-means-based pooling layer for convolutional neural networks that explicitly models local uncertainty and ambiguity. Conventional pooling operations, such as max and average, apply rigid aggregation and often discard fine-grained boundary information. In contrast, our method computes soft memberships within each receptive field and aggregates cluster-wise responses through membership-weighted pooling, thereby preserving informative structure while reducing dimensionality. Being differentiable, the proposed layer operates as standard two-dimensional pooling. We evaluate our approach across various CNN backbones and open datasets, including CIFAR-10/100, STL-10, LFW, and ImageNette, and further probe small training set restrictions on MNIST and Fashion-MNIST. In these settings, the proposed pooling consistently improves accuracy and weighted F1 over conventional baselines, with particularly strong gains when training data are scarce. Even with less than 1% of the training set, our method maintains reliable performance, indicating improved sample efficiency and robustness to noisy or ambiguous local patterns. Overall, integrating soft memberships into the pooling operator provides a practical and generalizable inductive bias that enhances robustness and generalization in modern CNN pipelines.

---

Keywords

---