Applied Research on the Automatic Detection of Lung Nodules ROI Based on Top-Hat and Gabor Filter

摘要

Computer aided diagnosis plays an important role in automatic detection of abnormal shadow area on CT images. The method of an automatic detection of lung nodules regions of interest (ROI) is presented in this paper. Firstly, the 1ung areas are segmented from CT images. Then, the initial ROI including nodules and blood vessels in lung areas are extracted by Top-hat filter. Thirdly, the second ROI are extracted by Gabor filter. Finally, compare the images to get more accurate lesion areas suspected nodules. Experimental results show that the method can extract the region of interest accurately and completely.