Character recognition method and device based on progressive coding and electronic equipment

Abstract

The invention discloses a character recognition method based on progressive coding, belongs to the technical field of computers, and is helpful for improving the accuracy of character recognition in an image. The character recognition method comprises the following steps: pre-training at least three cascaded coding modules and a coding-decoding network model of a decodinig module corresponding to each coding module, and in a prediction stage, sequentially executing the at least three coding modules to carry out progressive depth feature coding on a target image, thus respectively obtaining the coding output of each coding module; respectively executing each decoding module, and decoding the coding output of the corresponding coding module to obtain a character recognition result of each decoding module for the target image; and performing fusion processing on the text recognition result of the target image by each decoding module, and determining a character recognition result of the target image by the encoding and decoding network model, so that the recognition accuracy of characters in the image is improved.

Description

technical field [0001] The embodiments of the present application relate to the field of computer technology, and in particular, to a progressive encoding-based character recognition method, device, electronic equipment, and computer-readable storage medium. Background technique [0002] As a common technology in machine vision, text recognition is widely used in application scenarios such as document analysis, certificate recognition, street view recognition, and automatic information entry. In the prior art, mainstream methods for character recognition are all composed of encoding-decoding structures. Among them, the encoding process is to convert the input image information into a representation vector; the decoding process is to convert the representation vector generated by the encoding process into an output target sequence. In the prior art, the encoding process is usually realized by using a convolutional neural network, or a network structure of convolutional neura...

AI Technical Summary

Problems solved by technology

[0003] There are at least the following defects in the text recognition method in the prior art: 1, the encoding-decoding process method of convolutional neural network+serial connection classification decoding technology, only considers the visual feature of the image, if a certain character in the image is polluted (such as occlusion), the recognition accuracy will be greatly reduced; 2. The encoding-decoding process method of convolutional neural network + recurrent neural network + self-attention mechanism relies too much on context information, for example, if a certain word or If the frequency of occurrence of a sentence is too high, the degree of recognition of words similar to the word will cause it to be too sensitive to the distribution of training data, and the recognition accuracy will also be reduced.

Images