An Automated Recognition Technique for Lung Vessel Tree Using Chest Multi-slice CT Images

An Automated Recognition Technique for Lung Vessel Tree (ARTLVT) was proposed. ARTLVT was a one of 3-dimensional processing techniques, and had a recognition engine named Successive Region Growing (SRG). ARTLVT automatically found a cross-section of trachea from multi-slice CT images of hundreds. ARTLVT applied SRG to the cross-section and extracted trachea, bronchi and both lung fields. Based on the fact that lung vessels went into lung field together with bronchus through a narrow area (named window of lung field), ARTLVT found the window and applied SRG again for the lung vessels. SRG well recognized vessel branching point and all the branching point positions were stored in the lung vessel tree. ARTLVT was successfully applied to multi-temporal sets of actual multi-slice CT images.