Provided is an automatic segmentation system comprising an image segmentation unit for acquired images which segments a chest image in chest CT image through 3-dimensional region growing algorithm using 3-dimensional labeling； an image segmentation unit for lung and bronchus images which segments bronchus images based on morphologic characteristics of bronchi and variable threshold values of the chest image, determines if the thoracic wall is abnormal through the chest image and in case of abnormality, obtains a lung image through a modified rolling ball algorithm； and an image segmentation unit for a pulmonary vascular image which obtains a pulmonary vascular image from the lung image, wherein the image segmentation unit for a pulmonary vascular image eliminates the diaphragm and any noises from the pulmonary vascular image to obtain a pulmonary vascular image including pulmonary arteries. Also, provided is an automatic segmentation method comprising: a CT image acquisition step (S10) wherein a chest CT image is obtained through a CT scanner； a chest image acquisition step (S20) wherein a chest image including the lung and bronchus parts is obtained from the chest CT image obtained in the chest CT acquisition stage (S10)； a bronchus image segmentation stage (S30) wherein a bronchus image is obtained from the chest image obtained in the chest image acquisition step (S20)； a lung image segmentation stage (S40) wherein a lung image obtained from the chest image obtained in the chest image acquisition stage (S20)； and a pulmonary vascular image segmentation step (S50) wherein a pulmonary vascular image is obtained from the lung image obtained in the lung image segmentation stage (S40).본 발명은 흉부 CT영상에서 3차원 레이블링을 통한 3차원 영역성장법으로 흉부 영상을 분할하는 획득영상분할부； 흉부 영상에서 기관지의 형태학적 특성 및 가변 임계값을 통해 기관지 영상을 분할하며, 흉부 영상에서 폐의 흉벽에 이상이 있는지를 판단하고, 이상이 있는 경우에는 변형된 롤링 볼 알고리즘을 통해 폐 영상을 획득하는 폐·기관지영상분할부； 및 폐 영상에서 폐혈관 영상을 획득하는 폐혈관영상분할부；를 포함하되, 폐혈관영상분할부는 폐혈관 영상에서 횡격막 및 잡음을 제거하여 폐혈관을 포함하는 폐혈관 영상을 획득하는
