Khair Razlan Othman
Oxford Centre for Islamic Studies Scholarship (OCIS)
Automated Localization of Fetal Abdominal Ultrasound Structures
In the medical field and in particular in the field of fetal ultrasound, the quest of quality and certification of clinical decision-making has recently stimulated great interest among the healthcare practitioners. Diagnostics of fetal growth and abnormalities heavily depend on clinical judgment, which are translated into finding a good viewing plane from which the biometric parameters can be measured. Among the key criteria that define such a fetal ultrasound biometry plane are the presence and absence of particular anatomical landmarks in the image.
An efficient detection of an anatomical structure is a fundamental task and often serves as a prerequisite for the higher-level analysis and understanding of medical images. In particular, our initial focus is to automatically localize and categorize key anatomical structures of an abdominal ultrasound scan using machine learning technique that enable an accurate assessment of fetal biometric measurements.