Improving the performance of convolutional neural networks using a fuzzy logic based pooling method
9 viewsDOI:
https://doi.org/10.54939/1859-1043.j.mst.112.2026.29-37Keywords:
CNN; Deep learning; Fuzzy logic; Pooling layer; Classification.Abstract
In automatic image recognition and classification tasks, convolutional neural networks (CNN) are widely used, in which pooling operations play an important role in aggregating and highlighting image feature characteristics. This paper proposes a fuzzy logic based pooling method that integrates minimum, average, and maximum feature responses through a simple Takagi–Sugeno inference mechanism. Unlike conventional pooling methods, the proposed approach explicitly models uncertainty in feature aggregation, enabling a better balance between salient local features and contextual in order to preserve important image information. The effectiveness of the proposed method is evaluated on MNIST and CIFAR-10 datasets. Experimental results demonstrate that the proposed fuzzy pooling method achieves superior classification performance compared to conventional pooling methods, while maintaining comparable computational complexity. This enables the method to be effectively applied in CNN based image classification tasks.
References
[1]. R. Bhargava, N. Arivazhagan & K.S. Babu, "Hybrid RMDL-CNN for speech recognition from unclear speech signal". Int J Speech Technol 28, 195–217, (2025).
[2]. S. Chinnaiyan, A.B. Haffishthullah, S. Naveen, et al. "Optimized graph convolutional shunted self-attention neural network for multilingual speech-to-text training using cross-language voice conversion of speech representations". Int J Speech Technol 29, 22, (2026).
[3]. M. Bojarski et al., “End-to-end learning for self-driving cars,” arXiv preprint, arXiv:1604.07316, (2016).
[4]. A. G. Gad, “Particle swarm optimization algorithm and its applications: A systematic review,” Archives of Computational Methods in Engineering, Vol. 29, No. 5, pp. 2531–2561, (2022).
[5]. Q. Li et al., “EEG-based anxiety emotion classification using an optimized convolutional neural network and transformer,” Signal, Image and Video Processing, Vol. 19, pp. 501–509, (2025).
[6]. J. Zhang et al., “PCB defect recognition by image analysis using deep convolutional neural network,” Journal of Electronic Testing, Vol. 40, No. 5, pp. 657–667, (2024).
[7]. P. S. Geidarov, “Analytical calculation of weights of convolutional neural networks,” Optical Memory and Neural Networks, Vol. 33, No. 2, pp. 157–177, (2024).
[8]. X. Lin et al., “LIBS feature variable extraction method based on convolutional neural network,” Journal of Applied Spectroscopy, Vol. 92, No. 1, pp. 218–224, (2025).
[9]. C. Tang et al., “DeepSCNN: A simplicial convolutional neural network for deep learning,” Applied Intelligence, Vol. 55, pp. 281–295, (2025).
[10]. M. Zhu et al., “ShuffleNeXt: Modern lightweight convolutional neural network architecture,” Pattern Analysis and Applications, Vol. 27, No. 4, pp. 123–135, (2024).
[11]. C. Y. Lee, P. Gallagher, and Z. Tu, “Generalizing pooling functions in convolutional neural networks: Mixed, gated, and tree pooling,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 40, No. 4, pp. 863–875, (2017).
[12]. R. Nirthika, S. Manivannan, and A. Ramanan, “An experimental study on convolutional neural network-based pooling techniques for the classification of HEp-2 cell images,” Proc. of 10th International Conference on Information and Automation for Sustainability (ICIAfS), Colombo, Sri Lanka, (2021).
[13]. M. S. Greeshma and V. R. Bindu, “Single image super-resolution using fuzzy deep convolutional networks,” Proc. of International Conference on Technological Advancements in Power and Energy (TAP Energy), Kollam, India, (2017).
[14]. M. M. Hasan et al., “FP-CNN: Fuzzy pooling-based convolutional neural network for lung ultrasound image classification with explainable AI,” Computers in Biology and Medicine, Vol. 165, Article no. 107407, (2023).
[15]. S.B. Tharun, S. Jagatheswari, "A U-shaped CNN with type-2 fuzzy pooling layer and dynamical feature extraction for colorectal polyp applications". Eur. Phys. J. Spec. Top. 234, 2627–2635 (2025).
[16]. M. Sarkar, A. Mandal, "Atrous convolution and fuzzy pooling-based multi-channel CNN model for accurate follicle segmentation". Evolving Systems 17, 37 (2026).
[17]. C. Ozdemir, Y. Dogan, and Y. Kaya, “A new local pooling approach for convolutional neural networks: Local binary pattern pooling,” Multimedia Tools and Applications, Vol. 83, No. 12, pp. 34137–34151, (2023).
