Handwritten character-oriented one-stage automatic identification and translation method

Abstract

The invention discloses a handwritten character-oriented one-stage automatic identification and translation method. The method mainly comprises a text identification method and an end-to-end identification and translation method. According to the method, an attention mechanism is used for replacing an RNN structure in the CRNN, so that calculation can be parallelized, and the calculation cost is reduced; in the training process of the Transformer model, random replacement is carried out on input of a decoder, the situation of prediction errors in the prediction process is simulated, and the over-fitting problem is relieved; according to the method, an end-to-end recognition and translation model is provided, an end-to-end model is trained by using a transfer learning-based mode, a recognition result does not need to be given explicitly, and the picture content is directly translated. The method has the following advantages: 1, the training speed of the text recognition model is greatlyimproved; 2, the decoder input is randomly replaced in the training stage, so that the generalization ability of the recognition model is greatly improved; and 3, the translation accuracy of the end-to-end recognition and translation model is higher than that of the two-stage model.

Description

technical field [0001] The invention relates to a method for recognizing and translating handwritten characters in one language to another in a single stage. Background technique [0002] Current handwriting recognition and translation methods include the following methods: [0003] (1) Supervised deep learning method: first train the deep learning model on the training set, and use the trained model to classify when predicting new samples, such as figure 1 shown. figure 1 Above the dotted line is the training phase, and below it is the prediction phase. [0004] (2) Text line recognition technology: the system input is a picture containing a text line, and the input is a character string of the picture content, such as figure 2 shown. The most widely used in this field is the framework based on convolutional neural network + cyclic neural network. This method uses a combination of convolutional neural network (CNN) and cyclic neural network (RNN). The model is called C...

AI Technical Summary

Problems solved by technology

[0010] 2. Based on the sequence-to-sequence model of the Encoder-Decoder framework, the behavior of the decoder training phase and the prediction phase are inconsistent, which will lead to overfitting problems

[0011] 3. The recognition and translation system based on the two-stage method (recognition first and then translation) cannot avoid the impact of recognition errors on translation results

Images