Text image direction correction method and device

Abstract

The invention provides a text image direction correction method and device, and the method comprises the steps: intercepting a pixel block of a target printed single character from a target text imageaccording to a preset interception rule, and generating a printed character matrix corresponding to the pixel block; respectively obtaining a first dimension reduction vector corresponding to the printing word matrix and a second dimension reduction vector corresponding to the rotated printing word matrix by applying a preset principal component analysis model, wherein the principal component analysis model is an unsupervised machine learning model which is obtained by pre-training and is used for determining the characteristics of the printing word matrix; obtaining a target row vector corresponding to the first dimension reduction vector from a pre-obtained feature vector library, and obtaining an inclination angle and an inclination direction of a target text image according to the first dimension reduction vector, the second dimension reduction vector and the target row vector so as to correct the direction of the target text image. The accuracy and efficiency of recognizing the inclination angle of the text image can be improved, and then the accuracy and efficiency of text image direction correction are improved.

Description

technical field [0001] The present application relates to the technical field of data processing, in particular to a text image orientation correction method and device. Background technique [0002] In many machine vision scenarios, such as target recognition and text detection, the general model can better recognize text images with normal text orientation, that is, text images whose text orientation is parallel to the horizontal axis of the reference coordinate system, but for oblique angles For text image samples, the error rate of model recognition will increase significantly, so the direction of the text image needs to be corrected before inputting it into the model. [0003] Text orientation correction is generally divided into traditional image processing and deep learning. The more common method of traditional image processing is the Fourier transform method. First, the tilt angle of the image is checked in the frequency domain, and then corrected. This method can ...

AI Technical Summary

Problems solved by technology

The more common method of traditional image processing is the Fourier transform method. First, the tilt angle of the image is checked in the frequency domain, and then corrected. This method can only detect the relative value of the text rotation direction, and cannot identify the specific rotation. direction, so it is impossible to recognize text rotated 90 degrees and 270 degrees, or 0 degrees and 180 degrees; and deep learning is to recognize tilted pictures through neural network training, and to recognize pictures with a certain type of angle. This method requires a lot of data For training, and the unified data is only suitable for a single scene, and classify each angle at the same time, more types are required, the model is very complicated, and has no practical application value

Images