ROI based detection of abnormalities in lungs using medical image processing

Acquiring proper lung segmentation and detecting abnormalities in it is a tedious task for radiologists dealing with it. In this study, an approach to proper segmentation of the lungs is being tried to achieve and accompanied with detection of any abnormalities in it. The Segmenting part includes the creation of two masks manually each for lungs and the masks are iteratively layered to get the proper segmentation with active contour models. After getting the segmentation of lungs, an abnormal case is studied and abnormalities are detected based on the region of interest (ROI) by plotting the centroid and weighted centroid(center of mass) and calculating and comparing the standard deviations, abnormalities or any known structure can be detected instead of looking for the whole image. The bounding box is drawn for the structure if standard deviation goes beyond the calculated value.