Text recognition model training method, text recognition method, device and equipment

Abstract

The invention discloses a text recognition model training method, a text recognition method, device and equipment, and belongs to the technical field of image recognition, and the text recognition model training method comprises the steps: obtaining an image sample set, an image sample in the image sample set comprising a text image and a text label associated with the text image; performing sample expansion on the image sample set, and dividing the image sample set after sample expansion into a training set, a verification set and a test set; performing iterative training on a text recognition model according to the training set and the verification set, the text recognition model being constructed by replacing an original VGG network in a CRNN network model with an SE-ResNet network andsequentially cascading the SE-ResNet network with a BiLSTM network layer and an attention mechanism layer; and performing performance test on the text recognition model after iterative training according to the test set. According to the embodiment of the invention, the feature extraction capability of the text recognition model can be improved, and the feature vector decoding effect is improved,so that the text recognition accuracy is improved.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a text recognition model training method, text recognition method, device and equipment. Background technique [0002] The text recognition task requires certain image processing to recognize the text content in the image. Text recognition can be applied to many fields, such as sorting of letters and parcels, editing and proofreading of manuscripts, summary and analysis of a large number of statistical reports and cards, processing of bank checks, statistical summary of commodity invoices, identification of commodity codes, commodity warehouses management, as well as document retrieval, identification of various documents and office automation of financial bill processing, etc. It is convenient for users to quickly enter information and improve work efficiency in all walks of life. [0003] At present, there are two categories of text recognition methods. One is based...

AI Technical Summary

Problems solved by technology

[0005]Based on traditional image processing methods, it is often necessary to manually design some features, and some rules are interspersed in the middle to correct the improper processing of the algorithm, and the image background is complex , There is a lot of interference, and the characters are seriously bonded, the effect of the traditional method is not very good

Although the CRNN-based method uses deep learning to reduce manual intervention and improve the accuracy of text recognition and the stability of the algorithm, the original CRNN model uses the traditional VGG network for convolution to extract feature sequences and uses two-way LSTM. CTC is used to deal with the problem of converting variable length sequences into text, so for more complex and changeable text recognition tasks, there will be problems of weak model feature extraction, insufficient effective coding, and long-term and short-term dependencies, which will affect the recognition results.

Images