Image segmentation using multiple and partially evolved Hopfield neural networks
Ming Li, Xiaoqin Yang, Zhou Lin-xia · Society of Instrument and Control Engineers of Japan · 2003
This paper proposed an image segmentation method with multiple and partially evolved Hopfield neural networks. These neural networks, which are used to generate the threshold surface by interpolating the gray levels of the edge points into a segmentation threshold surface, are partially evolved to escape from possible local optima state.