Handwritten Chinese Character Recognition Based on Attention Mechanism

Handwritten Chinese character recognition is a key technology widely used in the fields of signature verification and paper document digitization. Because the strokes of Chinese characters are complex and varied, and the characters with similar shapes are difficult to distinguish, the traditional model often faces the problem of insufficient feature extraction when dealing with Chinese characters with different writing styles and cohesive strokes. To address this challenge, this study proposes a convolutional neural network model base on Improved Convolutional Block Attention Module,the feature information of different scales can be effectively fused.By adding the batch normalization layer to the model, the convergence speed of the model is accelerated,. The experimental results show that the improved model has more significant performance improvement than the traditional model when dealing with the above complicated situations