Deep Learning Applications for Finding Geometric Constraints of Chronic Subdural Hematomas

This thesis explores the use of deep learning in the medical field, specifically in the diagnosis and treatment of chronic subdural hematoma (cSDH). The research focuses on developing automated methods to identify geometric constraints in brain CT images, a critical step in standardizing and enhancing the treatment of cSDH. The study involves brain stripping techniques, ideal midplane detection, and the development of a slice selection algorithm. It utilizes the Design Science Research Methodology for a structured approach and evaluates its solutions through numerical results and visual inspections. The brain stripping method can successfully identify skull masks. The midplanes are detected with a 3.07-degree angle difference from the ground truth.