Handwritten Digital Image Recognition based on Fusion of Multiple Machine Vision Algorithms
B Lin, Zhi Yang, Minsheng Yang, Xiaole Xie, Xiong Wang, Huiting Liu, Yuntong Lan · 2023
Handwritten digit recognition technology is increasingly used in a wide range of fields, but the recognition rate still has room for further improvement. Currently, there are a large number of studies focusing on the use of activation functions to improve the handwritten digit recognition rate, compared with ReLU, Sigmod and other activation functions, this paper chooses to take the Mish activation function to optimise the handwritten digit recognition algorithm, using Pytorch machine learning framework, designed a deep convolutional neural network model based on the MNIST dataset to carry out digit recognition, and to improve the accuracy of model recognition through the learning of image features for digit classification to improve the model recognition accuracy. The experimental results show that the Mish activation function can increase the accuracy of handwritten digit recognition to 99%.