The techniques used to perform accurate static gesture recognition have been described in the literature, and the most commonly used methods are neural networks, support vector machines, and simple patterning techniques. Neural network is an information processing system developed as a mathematical model of human cognition and neurobiology. A neural network is characterized by connection patterns between neurons, methods to determine connection weights, and its activation function. Hasan proposes gesture recognition based on shape analysis. Six static gestures were tested by using a multilayer perceptual neural network and a back-propagation algorithm. The neural network structure consists of a hidden layer, an input layer and an output layer (6 nodes). They obtained a recognition rate of 86.38% on the training set of 30 images and the test set of 84 images. Xu et al. developed a virtual training system based on static gesture recognition and hand translation rotation. The input data was captured by data gloves with 18 sensors. In order to identify gestures, a feedforward neural network with 40 nodes hidden layer, input layer of 18 nodes and output layer of 15 nodes was used. The backpropagation algorithm was selected as the training method. The test performed on 300 gesture datasets from 5 individuals achieved 98% recognition performance. Static gesture recognition was done through self-growth and selforganized neural network. The algorithm tested 31 gestures derived from the combination of finger display and hidden. The data was collected from the camera and the background was unchanged, obtaining a recognition rate of 90.45%, but the average calculation time of CPU at 3 GHz was 1.5s. The algorithm is arranged in two steps, which reduces the difficulty of one-time arrangement and effectively solves the problem of Olympic schedule arrangement.