open access

Journal of Artificial Intelligence and Digital Health

P e e r - R e v i e w e d  
Bi-Annual (Two issues per year)

ISSN: 3139-6267 (Online)
DOI Prefix (Crossref): 10.67238

Automated Classification of Malignant Lymphoma Subtype using Autoencoder-Enhanced Convolutional Neural Networks
Research Article - Volume: 1, Issue: 2, 2026 (November)
Automated Classification of Malignant Lymphoma Subtype using Autoencoder-Enhanced Convolutional Neural Networks Download PDF

Bekele Mulat Enyew1*, Engidaw Awoke Muche1, Yonael Mulat Eneyew2, Ewunate Assaye Kassaw3, Ibrahim Gashaw4

1Department of Information Technology, College of Informatics, University of Gondar, Gondar, Ethiopia
2Department of Medicine, College of Medicine and Health Science, Yekatit 12 Hospital Medical College, Addis Ababa, Ethiopia
3Department of Biomedical Engineering, Institute of Technology, University of Gondar, Gondar, Ethiopia
4Department of Information Systems, College of Informatics, University of Gondar, Gondar, Ethiopia

*Correspondence to: Bekele Mulat Enyew, Department of Information Technology, College of Informatics, University of Gondar, Gondar, Ethiopia, E-mail:

Received: August 18, 2026; Manuscript No: JAID-26-5958; Editor Assigned: August 20, 2026; PreQc No: JAID-26-5958 (PQ); Reviewed: September 08, 2026; Revised: October 08, 2026; Accepted: October 12, 2026; Published: November 09, 2026

ABSTRACT

Background

Malignant lymphoma is a cancer originating in the lymphatic system and affecting various lymphatic tissues. With its increasing incidence, accurate reporting and analysis are becoming increasingly critical. The integration of artificial intelligence, particularly deep learning, has revolutionized cancer detection and prognosis. Automated histopathology image analysis using deep learning minimizes observer variability and reveals previously unnoticed features.

Methods

This research aims to enhance malignant lymphoma classification by improving model accuracy and image quality through preprocessing, autoencoders, and data augmentation. A convolutional neural network (CNN) was optimized using hyperparameter tuning for superior performance. The model was trained on a Kaggle dataset consisting of 3,740 PNG images (224x224x3), employing techniques such as histogram equalization, median filtering, and Gaussian filtering to enhance image quality and feature visibility.

Results

Autoencoders were used to refine the encoding and decoding processes, significantly improving classification accuracy. The optimal CNN model was trained for 100 epochs with a batch size of 32 and a learning rate of 0.001, achieving a training loss of 0.1884 and an accuracy of 99.2%. During validation, the model attained a loss of 0.4271 and an accuracy of 97.7%. On test data, the model achieved a loss of 0.3820 and an accuracy of 98%. Additionally, a web-based tool developed using Flask and Python demonstrated a classification probability match of 98.81%.

Conclusion

This study highlights the potential of CNN-based approaches to enhance lymphoma diagnosis accuracy and efficiency, thereby alleviating the workload on radiologists.

Keywords: Malignant Lymphoma; Deep Learning; Convolutional Neural Network; Image Preprocessing; Autoencoders

Citation: Enyew BM, Muche EA, Eneyew YM, Kassaw EA, Gashaw I (2026). Automated Classification of Malignant Lymphoma Subtype using Autoencoder-Enhanced Convolutional Neural Networks. J. Artif. Intell. Digit. Health. Vol.1 Iss.2, November (2026), pp:189-198.
Copyright: © 2026 Bekele Mulat Enyew, Engidaw Awoke Muche, Yonael Mulat Eneyew, Ewunate Assaye Kassaw, Ibrahim Gashaw. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.