Knowledge-based method for segmentation and analysis of lung boundaries inchest X-ray images

We present a knowledge-based approach to segmentation and analysis of the l ung boundaries in chest X-rays. Image edges are matched to an anatomical mo del of the lung boundary using parametric features. A modular system archit ecture was developed which incorporates the model, image processing routine s, an inference engine and a blackboard. Edges associated with the lung bou ndary are automatically identified and abnormal features are reported. In p reliminary testing on 14 images for a set of 18 detectable abnormalities, t he system showed a sensitivity of 88% and a specificity of 95% when compare d with assessment by an experienced radiologist. (C) 1999 Elsevier Science Ltd. All rights reserved.