ABSTRACT: This work has employed a deep learning strategy to automatically detect and classify white blood cell (WBC) cancers, including leukemia, using histopathology images. This technique analyses histopathological images and uses convolutional neural networks (CNNs) to accurately detect distinct WBC cancer subtypes. When compared to pathologists who manually read cases, our method provides answers more quickly and accurately. Tested extensively on multiple datasets, our method consistently outperforms existing methods in terms of sensitivity, specificity, and overall accuracy. The study has also been made to improve the effectiveness of transfer learning techniques, which allow our model to adapt and perform well on different datasets. Because of its versatility, it can be applied in real-world clinical settings, which has the potential to revolutionize personalized medicine approaches to WBC cancer diagnosis and treatment. Additionally, our method employs explainable AI techniques to give doctors greater assurance and understanding by revealing the model's decision-making process. More informed treatment decisions by healthcare professionals lead to better outcomes for patients with WBC malignancies. By combining advanced deep learning methods with interpretable models, our research provides a significant step toward integrating AI-driven treatments into standard clinical practice. This has the potential to significantly improve patient care and outcomes in the field of oncology.
KEYWORDS: White Blood Cancer Detection, Artificial intelligence, Deep learning, Histopathological Images, Convolutional Neural Networks (CNNs), Benign, Malignant, Rank-Based Ensemble, Inceptionv3, Xception, MobileNet
|
DOI:
10.17148/IJARCCE.2024.13383
