Study on the interactive serial medical images segmentation

Interactive segmentation algorithm, which can provide accurate, reproducible boundary extraction, is widely applied to medical images. For hundreds of serial CT images of a patient, it is intolerable for radiation therapy physicians to delineate cancer target and some key normal tissues interactively one by one. The paper suggests a method to segment key layers of the image sequence with interactive live-wire segmentation algorithm by doctors firstly, and then with the continuity between layer images, snake active contour is used to get object contours on the large number of middle layer images by the interpolating between adjacent key layers. For the interactive live-wire segmentation algorithm, heap sort is used for searching the globally optimal path from the start node to the goal node to reduce the complexity of the calculation time. Algorithm analysis and experiments indicate that the presented methods can evidently improve the efficiency of segmentation of serial medical images.