An adaptive segmentation technique for clinical endoscopic image processing

Computerized segmentation techniques have been successfully employed in different modalities of medical imaging, particularly in magnetic resonance. Segmentation of clinical endoscopic images for intelligent endoscopy has been attempted in this research. An adaptive lumen center detection algorithm is developed and a new segmentation method is proposed in which a splitting-merging based algorithm is combined with the lumen center detection. The results show that endoscopic images can be segmented for identification of different regions of clinical interest using the proposed algorithm, thus contributing to intelligent endoscopy.