A LOCAL CLUSTERING SEGMENTING METHOD FOR ANATOMICAL CROSS-SECTIONAL IMAGE OF CHINESE DIGITAL HUMAN / 解剖学报

ABSTRACT

Objective This article aimed to describe an interactive segmenting method for serial anatomical cross-sectional image of Chinese Digital Human(CDH) using commercially available software,Photoshop CS(Adobe;San Jose,CA,Corel KnockOut 2.0 plug-in) and Matlab 7.0(Math Works;Worcester,MA). Methods First,the interesting region was segmented from the image interactively utilizing the Knochout menu of Adobe Photoshop which had a powerful masking function.And then smooth contour was acquired precisely by morphological operation function and edge detection operator(the Sobel method) in Matlab. Results The segmentation and classification of bone,muscle and main vascular nerve were fulfilled successfully and the smooth contour lines were acquired with precise details and accurate localization.Conclusion This local clustering segmenting method might segment and classify the colored image of Chinese Digital Human quickly and precisely.