中文说明：
极限学习机是由黄广斌在2005年作为一种新的单隐层前馈神经网络提出的，具有与神经网络（NN）相同的全局逼近性质，且其参数学习无需迭代，速度明显快于现有的神经网络。目前在岩性识别、LF终点温度软测量、穿孔机导盘转速测量、软测量建模、图像识别等方面有所应用，但将其用于图像分割中的应用还较少。
English Description：
Extreme learning machine (ELM) was proposed by Huang guangbin as a new single hidden layer feedforward neural network in 2005. It has the same global approximation property as neural network (NN), and its parameter learning does not need iteration, and its speed is obviously faster than the existing neural network. At present, it has been applied in lithology identification, lf end point temperature soft measurement, piercer guide disc speed measurement, soft sensor modeling, image recognition and so on, but its application in image segmentation is still less.