Improved LeNet-5 Model Based On Handwritten Numeral Recognition
Shuai Tan, Zhiyi Tan · 2019
Handwritten numeral recognition is a branch of character recognition, it is widely used in large-scale financial statistics, bank checks, financial statements and the like. In this paper, the traditional LeNet-5 model is improved, and the 1*1 convolutional layer in LeNet-5 is removed, and the moving average model is added. The experimental results on the MNIST dataset show that the improved model can identify handwritten numeral more effectively, and the recognition rate reaches 99.06%.