Bone Marrow Classification using Deep Learning

• Author(s): Disha N ; Mahesh N ; Rashmi C R
• Paper ID: 1710929
• Page: 1661-1667
• Published Date: 30-09-2025
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 9 Issue 3 September-2025

Abstract

The diagnosis of hematological disorders such as leukemia, lymphoma, and multiple myeloma often requires accurate and timely classification of bone marrow cells. Traditionally, this task is carried out manually by pathologists, who visually inspect bone marrow smear images under a microscope. While effective, this manual process is time-consuming, prone to fatigue, and subject to inter-observer variability, which can impact diagnostic consistency. With the advent of Artificial Intelligence (AI) and particularly deep learning, there is an emerging opportunity to automate this critical step in clinical diagnostics. This project presents an automated framework for Bone Marrow Cell Classification using Deep Learning, designed to assist pathologists by reducing manual workload and improving diagnostic accuracy. The system follows a modular pipeline that begins with data acquisition from publicly available bone marrow smear datasets, followed by preprocessing techniques such as image resizing, normalization, and augmentation to improve generalization. Visualization of the dataset provides insights into class distribution and morphology. For classification, ResNet 50 is implemented, and transfer learning approaches with architectures such as ResNet and VGG are explored to enhance performance on limited medical datasets. The trained models are stored as .h5 files for reusability and deployment, while x`the prediction module processes new images and outputs class labels with The system is evaluated using clinically relevant metrics, including accuracy, precision, recall, F1-score, confusion matrix, and ROC curves, to ensure reliability in medical applications. Results demonstrate that the proposed approach achieves high accuracy and consistency compared to manual inspection, highlighting its potential as a decision-support tool in hematology.

Keywords

Bone Marrow Cell Classification, Hematological Disorders, Leukemia Detection, Medical Image Analysis, Deep Learning, Convolutional Neural Networks (CNN), ResNet50, Transfer Learning, Image Preprocessing (Resizing, Normalization, Augmentation), Computer-Aided Diagnosis (CAD), Artificial Intelligence in Healthcare, Medical Imaging Datasets, Model Evaluation Metrics (Accuracy, Precision, Recall, F1-score, Confusion Matrix, ROC-AUC), Pathology Automation, Clinical Decision Support Systems, Explainable AI (XAI) in Medical Imaging, Vision Transformers (ViT), Data Augmentation for Medical Images, Healthcare Automation, Diagnostic Consistency.

Citations

IRE Journals:
Disha N , Mahesh N , Rashmi C R "Bone Marrow Classification using Deep Learning " Iconic Research And Engineering Journals Volume 9 Issue 3 2025 Page 1661-1667

IEEE:
Disha N , Mahesh N , Rashmi C R "Bone Marrow Classification using Deep Learning " Iconic Research And Engineering Journals, 9(3)