激光共焦显微生物医学图象分割方法的研究

提出了用于激光共焦显微生物医学图象分割的循环递增阈值分割法。先用最低分割阈值把图象分割成子区域，然后测量每个子区域的灰度均匀度。如果灰度均匀度不满足灰度均匀度分割准则，则增加分割阈值，对该区域进行再分割。循环这个过程，直到没有区域需要分割为止。介绍了围线积分法区域标号和灰度均匀度分割准则的设计。给出了用该法分割图象的两个实例。

A STUDY OF THE SEGMENTATION FOR LASER SCANNING CONFOCAL MICROSCOPIC BIOMEDICAL IMAGES

A recursively increasing threshold method for the segmentation of laser scanning confocal microscopic (LSCM) biomedical images is presented. In this method, an image is first segmented into sub regions in a lowest threshold. Then the intensity homogeneity of each sub region is examined. If the intensity homogeneity is not satisfied with a segmentation criterion, it is re segmented by increasing threshold. This process is iterated till no region needs resegmented. The advantage of this method is that the objects with wide intensity ranges and different intensity homogeneity can be segmented properly. The recursively increasing threshold method is first proposed. Then the method of contour integration region labeling and the design of intensity homogeneity segmentation criterion are presented. Two examples of image segmentation by this method are given at the end.