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 (AI) has emerged as a transformative approach in medical imaging, particularly for diagnosing hematological malignancies such as leukemia. White blood cancer originates in bone marrow and involves abnormal proliferation of white blood cells, requiring precise segmentation and classification for early detection and effective treatment. Traditional diagnostic methods based on manual microscopic examination are time-consuming, subjective, and prone to variability among observers. Recent advancements in deep learning, including convolutional neural networks (CNNs), attention-based models, and hybrid architectures, have significantly improved automated leukemia detection using bone marrow images. This review highlights AI techniques for segmentation and classification, with a focus on Deep Kronecker Neural Networks (DKNN), which enable efficient parameterization and enhanced feature representation for high-dimensional data. Various approaches, including CNN-based segmentation, transformer-based classification, and multimodal learning, demonstrate high accuracy in detecting leukemia subtypes. Despite these advancements, challenges such as data scarcity, class imbalance, domain adaptation, and interpretability remain. Emerging solutions such as generative adversarial networks, transfer learning, and attention mechanisms show promise in addressing these issues. Overall, advanced AI models offer significant potential for improving diagnostic accuracy and supporting real-time clinical decision-making.