IMAGE SEGMENTATION VIA ADAPTIVE K-MEAN CLUSTERING AND KNOWLEDGE-BASEDMORPHOLOGICAL OPERATIONS WITH BIOMEDICAL APPLICATIONS

Image segmentation remains one of the major challenges in image analys is, since image analysis tasks are often constrained by how web previo us segmentation is accomplished. In particular, many existing image se gmentation algorithms fail to provide satisfactory results when the bo undaries of the desired objects are not clearly defined by the image-i ntensity information. In medical applications, skilled operators are u sually employed to extract the desired regions that may be anatomicall y separate but statistically indistinguishable. Such manual processing is subject to operator errors and biases, is extremely time consuming , and has poor reproducibility. We propose a robust algorithm for the segmentation of three-dimensional (3-D) image data based on a novel co mbination of adaptive K-mean clustering and knowledge-based morphologi cal operations. The proposed adaptive K-mean clustering algorithm is c apable of segmenting the regions of smoothly varying intensity distrib utions. Spatial constraints are incorporated in the clustering algorit hm through the modeling of the regions by Gibbs random fields, Knowled ge-based morphological operations are then applied to the segmented re gions to identify the desired regions according to the a priori anatom ical knowledge of the region-of-interest. This proposed technique has been successfully applied to a sequence of cardiac CT volumetric image s to generate the volumes of left ventricle chambers at 16 consecutive temporal frames. Our final segmentation results compare favorably wit h the results obtained using manual outlining. Extensions of this appr oach to other applications can be readily made when a priori knowledge of a given object is available.