LukeNet: A lightweight CNN integrated with an XAI model for Smart acute lymphoblastic leukemia detection and management

Lightweight AI design enables real-world medical deployment: LukeNet achieves 99% accuracy while using significantly fewer computational resources than standard AI models, making it suitable for deployment on wearable and edge devices that operate under power and memory constraints—a critical requirement for practical healthcare applications rather than theoretical research.

Explainable AI improves clinical trustworthiness: The integration of XAI (explainable artificial intelligence) methods into LukeNet allows clinicians to understand why the system makes specific detection decisions, addressing a key barrier to AI adoption in healthcare where doctors need transparency to validate and trust automated recommendations.

Rigorous validation across multiple datasets strengthens reliability: The model was trained and tested on three different ALL datasets with five-fold cross-validation, and outperformed six established AI models (DenseNet121, MobileNet, ResNet50, etc.), suggesting the results are robust rather than overfitted to a single dataset.

Real-time patient monitoring becomes feasible: By combining lightweight AI with Internet of Medical Things (IoMT) wearables that track vital signs like temperature, heart rate, and oxygen saturation, the system enables continuous monitoring and early detection of acute conditions, potentially allowing timely clinical intervention outside traditional hospital settings.

Responsible AI requires balancing performance with accessibility: The research demonstrates that achieving high accuracy (99%) doesn't require deploying resource-intensive models; instead, thoughtful design can create efficient systems that are both accurate and deployable in resource-limited healthcare environments.