Artificial Intelligence Techniques for Segmentation and Classification of White Blood Cancer Cells in Bone Marrow Microscopic Images Using Deep Kronecker Neural Networks: Trends and Challenges

Abstract

Artificial Intelligence has emerged as a transformative paradigm in medical imaging, particularly in the diagnosis of hematological malignancies such as leukemia, a form of white blood cancer originating in the bone marrow and characterized by abnormal proliferation of white blood cells. Accurate segmentation and classification of these cells are essential for early detection and effective treatment. Traditional diagnostic approaches rely on manual microscopic examination, which is time-consuming, subjective, and prone to variability among experts. Recent advancements in deep learning, including convolutional neural networks, attention-based models, and hybrid architectures, have significantly improved automated leukemia detection using bone marrow images. This review focuses on advanced AI techniques for segmentation and classification, with particular emphasis on Deep Kronecker Neural Networks, which offer efficient parameterization and enhanced feature representation for high-dimensional medical data. The study examines modern approaches such as CNN-based segmentation, transformer-based classification, and multimodal learning frameworks that achieve high accuracy in detecting leukemia subtypes like Acute Lymphoblastic Leukemia and Acute Myeloid Leukemia. It also discusses key challenges including data scarcity, class imbalance, and model interpretability, while highlighting emerging solutions such as transfer learning, generative models, and attention mechanisms. Overall, the integration of advanced deep learning techniques demonstrates strong potential for improving diagnostic accuracy and supporting clinical decision-making in leukemia detection.