Research on handwritten score recognition algorithm of test papers based on improved convolution neural network
Xiyue Wang · 2022
The traditional way of marking papers relies mainly on manual work, and its process includes: marking, awarding marks, statistics, sorting and other processes. Teachers not only need to spend a lot of time and energy to increase the marking workload, but also prone to score summary error phenomenon. Nowadays, handwritten numeral recognition is gradually applied to various fields, which plays a huge role in information processing. This paper first analyzes the characteristics of common algorithms and builds an improved convolutional neural network to improve its recognition accuracy. Let it carry out deep learning of the handwritten scores on the test paper, which can be used for automatic recognition of the handwritten scores on the test paper. In the marking process, teachers only need to complete the grading work, and then the computer automatically completes the entry, summary and ranking of the scores. In this way, teachers can reduce the workload of marking, so that teachers can put their time and energy into teaching work.