← Back to Publications List

CBC-ViT: Blood Cell Type Classification Using a Hybrid CNN-Vision Transformer with Explainable AI

Students & Supervisors

Student Authors
Mahin Montasir Afif
Bachelor of Science in Computer Science & Engineering, FST
Abdullah Al Noman
Bachelor of Science in Computer Science & Engineering, FST
K. M. Tahsin Kabir
Bachelor of Science in Computer Science & Engineering, FST
Koushik Biswas Arko
Bachelor of Science in Computer Science & Engineering, FST
Supervisors
Md. Mortuza Ahmmed
Associate Professor, Faculty, FST

Abstract

Accurate classification of blood cell types plays a pivotal role in automated hematological analysis and serves as a fundamental step toward early blood-related disorder detection and diagnosis. In this study, we propose CBC-ViT, a novel hybrid Convolutional Vision Transformer model designed to classify five distinct blood cell types: basophil, erythroblast, monocyte, myeloblast, and neutrophil. The proposed architecture integrates a Convolutional Neural Network (CNN) backbone for hierarchical feature extraction with a Vision Transformer (ViT) encoder employing multi-head self-attention to capture long-range dependencies and contextual information. The model is optimized using the Adam optimizer with carefully tuned learning parameters to ensure stable convergence. Experimental results demonstrate the superior performance of CBC-ViT, achieving a test accuracy of 99.20%, with precision, recall, and F1-scores all exceeding 99%, thereby highlighting the robustness and reliability of the proposed approach. To further ensure transparency and trustworthiness, we incorporated Explainable AI (XAI) techniques, including Grad-CAM, Integrated Gradients, and SmoothGrad, which provide interpretable visual explanations of the model’s predictions, thereby assisting medical practitioners in validating automated outputs. The results underscore the potential of CBC-ViT as a reliable and explainable decision-support tool in hematological diagnostics.

Keywords

Blood cell hematological analysis Vision Transformer Convolutional Neural Network Explainable AI

Publication Details

  • Type of Publication:
  • Conference Name: IEEE International Conference on Biomedical Engineering, Computer and Information Technology for Health 2025 (BECITHCON 2025)
  • Date of Conference: 29/11/2025 - 29/11/2025
  • Venue: Eastern University, Dhaka, Bangladesh
  • Organizer: IEEE Bangladesh section and IEEE Engineering in Medicine and Biology Society Bangladesh
View on Publisher