基于双重边缘检测和区域生长的大脑皮层分割算法

A segmentation algorithm of cerebral cortex based on dual edge detections and regional growing

目的设计一种数字人脑切片图像自动分割算法,以实现对大脑皮层的准确分割。方法采用RGB空间彩色边缘检测与Canny算子边缘检测,联合种子点自动选择的区域生长算法,提取并分割大脑皮层,并与手工分割结果进行对比分析。结果自动分割的大脑皮层边缘完整、清晰、平滑,与手工分割结果的吻合度较高。结论采用该方法可较为准确地分割数字人脑切片图像中的大脑皮层。

Objective To design a automatic segmentation algorithm for digital cross-section slice images of human brain, in order to segment the cerebral cortex accurately. Methods Cerebral cortex was obtained and extracted from gray matter using RGB space edge detection, Canny operator detection and regional growing algorithm. A comparative analysis between the segmentation results of automatic algorithm and manual method was performed. Results Highly coincident to that segmented by manual method, cere- bral cortex segmented by automatic algorithm was characterized by complete, clear and smooth edge. Conclusion This algorithm is suitable for the segmentation of cerebral cortex in digital cross-section slice images.