论文部分内容阅读
将局部兴奋全局抑制振荡网络（ＬｏｃａｌｙＥｘｃｉｔａｔｏｒｙＧｌｏｂａｌｙＩｎｈｉｂｉｔｏｒｙＯｓｃｉｌａｔｏｒＮｅｔ－ｗｏｒｋ，简称ＬＥＧＩＯＮ）应用于图像分割．将侧电势引入振荡子的动态行为，有效地克服了噪声对主要区域的影响．用含噪声的二值图模拟了ＬＥＧＩＯＮ的时间演化．根据大脑区域分割特征聚合原理的相近性、相似性和连通性原则，设计了分割灰度图的网络连接权．用ＨＳＩ色空间设计了突出Ｈ分量的连接权，得到了比基于ＲＧＢ色空间分割更为满意的分割效果．
Locally Excited Globally Shifted Oscillator Net-work (LEGION) is applied to image segmentation. The introduction of side potential into the dynamic behavior of the oscillator, effectively overcome the noise impact on the main area. The time evolution of LEGION is simulated with the noisy binary map. According to the principle of similarity, similarity and connectivity of the brain region segmentation feature aggregation principle, the network connection right of the segmentation grayscale is designed. The HSI color space is used to design the connection weights that highlight the H component, and the segmentation result is more satisfied than that based on the RGB color space.