Design And Implementation Of Airway Tree Segmentation Algorithm Based On MSCT Images

Chronic Obstructive Pulmonary Disease (COPD) is one of the largest numbers of respiratory diseases, which not only influences patients’ health and quality of life seriously, but also causes a significant financial burden to the families and the society. With the emergence and development of Multi-slice Spiral CT (MSCT) technology, quantitative assessment based on MSCT images has been a powerful tool to evaluate the changed grade of tissues, study pathogenesis long-termly and follow treatment results. To segment airway tree timely and precisely has been the core technology and urgent need to address from massive MSCT images in COPD quantitative assessment system.On the basis of the above contexts, airway tree segmentation and centerline extraction algorithms were designed in the thesis referred in aided diagnosis of lung function analysis, with MSCT images as data resource and lung airway tree as main study object.Initially, related background knowledge was introduced in the fields of airway tree anatomical and iconology priori knowledge, medical image segmentation, centerline extraction and moving target tracking. Secondly, an automated algorithm was designed and implemented in the thesis based on otsu and multi-scale 3D region growing method to acquire airway tree trunk data; besides, airway tree centerline extraction was accomplished based on topology thinning algorithm, as well as centerline topological structure according to SWC file standard and branch directions fitted by least square method; finally, target tracking technology was drawn into airway segmentation in the thesis, and airway ending segmentation algorithm was designed and implemented based on Kalman filtering. Experiments testify the algorithms proposed possess desired accuracy and speed, and meet requirements in clinical aided diagnosis.