Semiautomatic Segmentation of Ventilated Airspaces in Healthy and Asthmatic Subjects Using Hyperpolarized MRI
Semiautomatic trachea removal. The example shown here is from an asthmatic subject. A corresponding HP 3He MRI image (a) and a binary ventilation image (b) are displayed in which a user-selected bounding box captures the trachea (c). An initial seed point is selected within the trachea, and a region-growing algorithm is applied to yield a resultant binary extraction (d). Thereafter, each element within the image space is labeled based on connectivity (e), and an area filter is used to isolate the trachea (f). A simple binary subtraction between the isolated trachea is then used (g) followed by a binary erosion (h). Then, each element within the image space is labeled again based on connectivity to isolate the right and the left lungs (i), and a binary dilation is thereafter applied (j)-(k) with a corresponding HP 3He MRI showing the outlined boundaries (l).