METHOD FOR SEGMENTING CHEST CT IMAGE DATA USING AN ANATOMICAL MODEL -PRELIMINARY-RESULTS

We present an automated, knowledge-based method for segmenting chest c omputed tomography (CT) datasets, Anatomical knowledge including expec ted volume, shape, relative position, and X-ray attenuation of organs provides feature constraints that guide the segmentation process. Know ledge is represented at a high level using an explicit anatomical mode l, The model is stored in a frame-based semantic network and anatomica l variability is incorporated using fuzzy sets, A blackboard architect ure permits the data representation and processing algorithms in the m odel domain to be independent of those in the image domain, Knowledge- constrained segmentation routines extract contiguous three-dimensional (3-D) sets of voxels, and their feature-space representations are pos ted on the blackboard, An inference engine uses fuzzy logic to match i mage to model objects based on the feature constraints, Strict separat ion of model and image domains allows for systematic extension of the knowledge base, In preliminary experiments, the method has been applie d to a small number of thoracic CT datasets. Based on subjective visua l assessment by experienced thoracic radiologists, basic anatomic stru ctures such as the lungs, central tracheobronchial tree, chest wall, a nd mediastinum were successfully segmented, To demonstrate the extensi bility of the system, knowledge was added to represent the more comple x anatomy of lung lesions in contact with vessels or the chest wall, V isual inspection of these segmented lesions was also favorable, These preliminary results suggest that use of expert knowledge provides an i ncreased level of automation compared with low-level segmentation tech niques, Moreover, the knowledge-based approach may better discriminate between structures of similar attenuation and anatomic contiguity. Fu rther validation is required.