This thesis introduces a set of algorithms aimed at analyzing human lungs through CT images. Specifically, for bronchoscopy procedures, the airway labeling algorithm ensures accurate and efficient navigation for clinicians. Moreover, the lobe segmentation algorithm facilitates clinicians' comprehension of lung disease distribution across different lobes. Furthermore, the thesis emphasizes the importance of visualizing the proposed algorithms, enabling clinicians to better understand the algorithms' predictions.
The algorithms developed in this thesis are based on the latest advances in artificial intelligence. In particular, deep learning methods are explored and crafted for CT image analysis. In conclusion, this thesis represents a noteworthy breakthrough in COPD imaging analysis, offering practical and efficient solutions that hold the potential to enhance patient care.
Weiyi Xie behaalde zijn masterdiploma aan de Tampere University of Technology in 2014. Zijn MSc-werk was gericht op het ondersteunen van vectormachine learning en het ophalen van op inhoud gebaseerde afbeeldingen. Xie begon in 2015 te werken aan verschillende industriële projecten met betrekking tot CT-analyse, met de nadruk op detectie van longknobbels. Xie ontwikkelt momenteel geavanceerde kwantitatieve CT-analysemethoden voor COPD-behandelplanning.