An Intelligent Approach to Diagnosing Severe Diseases

Early diagnosis of severe diseases and neurological disorders remains a significant challenge, primarily because the initial symptoms are often subtle and difficult to detect. Delayed onset can lead to serious consequences, such as the late initiation of treatment and a deterioration in patient outcomes. However, recent advancements in artificial intelligence (AI) and machine learning offer promising solutions to these issues. This thesis aims to study and develop AI-based solutions to address the challenges of diagnosing these diseases. The research focuses on two specific cases: Leukemia, a severe illness, and Autism, a neurological disorder. For Leukemia, we conducted a preliminary experiment using a deep learning system to assist in identifying malignant white blood cells from microscopic images. Additionally, to study Autism, we initiated an initial experiment using magnetic resonance imaging (MRI) of the brain to enhance the precision of intelligent diagnostic models. A second experiment explored the unique characteristics of human voice patterns to develop a more effective Autism diagnosis approach. This research is innovative in its methodology and the data used. The results are promising, highlighting the potential of AI to significantly improve diagnostic accuracy. Furthermore, the developed models can be integrated into applications to help healthcare professionals make more accurate and timely diagnoses