Fuhrmann G
Biol Cybern. 1981;40(1):39-48. doi: 10.1007/BF00326679.
"Modulo system" is a general information extracting and processing network for modelling the nervous system. Its essential features are briefly summarized. The concept is applied to visual pattern recognition. The input is a two-dimensional picture of an object. The network extracts the contours, the line-segments, the lines of different directions, and the vertices of different numbers of lines, in successive functional layers. Afterwards, it recognized the simplest known pattern which has the actually extracted features. Uncertainty, association, and tolerance are also concerned. The network is detailed for the classification of of three-dimensional geometrical objects.
“模系统”是一种用于对神经系统进行建模的通用信息提取和处理网络。简要总结了其基本特征。该概念应用于视觉模式识别。输入是物体的二维图片。网络在连续的功能层中提取轮廓、线段、不同方向的线以及不同线数的顶点。之后,它识别出具有实际提取特征的最简单已知模式。还涉及不确定性、关联和容差。针对三维几何物体的分类对该网络进行了详细阐述。