Fully Automatic Segmentation and Detection of Pulmonary Artery and Embolism in CTA images

摘要

Computer Aided Detection (CAD) system for detection of pulmonary embolism (PE) is proposed in CT angiography images. Our approach consists of three stages: firstly the pulmonary artery (PA) tree is extracted from the thorax in order to reduce the search area aiming to reduce the number of false positives. In the second step, PE candidates are detected within the segmented PA by using several intensity and shape based features which represent different properties of PE. In the third step, a filtering method is used to exclude false positive detection associated with the partial volume effect on the artery boundary, lymphoid tissue and noise and motion artifacts. The method was tested on 55 data scans (20 training data scans and 35 additional data scans for evaluation containing a total of 195 emboli). Resulting performance gave 94% detection sensitivity with an average 4.1 false positive detections per scan.