摘要：数字识别技术的研究不仅可以解决当下面临的数字识别问题，同时在图片

Abstract: The research of digital recognition technology can not only solve the current digital recognition problems, but also pave the way for image recognition and machine learning. This article mainly uses the convolutional neural network (CNN) method to realize handwritten digital picture recognition. First design the network model, then use the MNIST data set to train, and test the recognition accuracy of the network model, and then preprocess the handwritten digits and input them into the model to verify the correctness.

Abstract: The research of digital recognition technology can not only solve the problem of digital recognition, but also pave the way for image recognition, machine learning and so on. This paper mainly realizes handwritten digital image recognition by the covon neural network (CNN) method. First design the network model, then use MNIST data set training, and test the accuracy of the identification of the network model, and then the opponent writes numbers for the pre-processing of pictures, input into the model, verify the correctness.

Absrtact: the research of digital recognition technology can not only solve the problem of digital recognition, but also pave the way for image recognition and machine learning. This paper mainly uses convolutional neural network (CNN) to realize handwritten digital image recognition. Firstly, the network model is designed and trained with MNIST data set, and the recognition accuracy of the network model is tested. Then, the handwritten digits are preprocessed and input into the model to verify the accuracy.<br>