End-to-end printed Mongolian recognition translation method based on spatial transformation network

Abstract

The invention discloses an end-to-end printed Mongolian recognition translation method based on a spatial transformation network. Starting from the characteristics of Mongolian characters, a four-stage character recognition network is used for recognizing the Mongolian characters, and in the spatial transformation stage, the spatial transformation network is used for standardizing Mongolian character pictures; text features are extracted by using a CBAM-GRCNN in combination with context information, and sequence modeling is carried out by using a Morifier LSTM. In the prediction stage, a focusattention mechanism is used for solving the problem of attention drift, and prediction is carried out in combination with a GRU network; in the translation process, a Reformer model is adopted, the Reformer model changes a traditional multi-head attention mechanism into an attention mechanism based on local sensitive hash, a traditional residual network is replaced with a reversible residual network, and the feedforward network is blocked, so that the time and space complexity of the model is reduced, and the problems of insufficient memory and low speed of training long-sequence data are relieved.

Description

technical field [0001] The invention belongs to the technical field of optical character recognition (OCR) and machine translation, in particular to a method for end-to-end printed Mongolian recognition and translation based on a space transformation network. Background technique [0002] Optical character recognition is a technology that uses computers to extract text from pictures into text. It is one of the most effective means to solve the problem of converting pictures to text. Machine translation can convert a language into the target language, which is an effective solution to language barriers. Way. With the development of deep learning, the use of deep learning for optical character recognition and machine translation tasks has become the mainstream. Google, Baidu, Youdao, etc. have conducted a lot of research on optical character recognition and machine translation, and have developed practical application. [0003] Before the emergence of end-to-end character re...

AI Technical Summary

Problems solved by technology

[0005] At present, optical character recognition has matured in mainstream languages, and has a good recognition rate for distorted characters, but there are few studies on small languages, especially the serious lack of databases, and the recognition and translation have not achieved good results.

Images