A general OCR training data generation system and method based on machine learning

Abstract

The invention discloses a general OCR training data generation system and method based on machine learning. The method is implemented based on the system, and the steps include randomly extracting 5-10 characters from a corpus as text information; randomly selecting from a font library The font generates font information; the background image is randomly selected from the image library, and the image is cut according to the text information generated by the font information; the pixel RGB value of the image background is analyzed by a clustering algorithm to find the cluster center, and then the image is selected from the text color library Randomly extract colors, calculate the distance from each color to the RGB value of the cluster center of the background color value, and then randomly select the text color from the farthest color; combine text information, font information, background image, and text color to generate and directly use Based on the pictures trained by the text recognition model, the method uses real scene background pictures, adds font colors through cluster analysis, and realizes the completely automatic simulation generation of the real training pictures of the text recognition model.

Description

technical field [0001] The invention belongs to the field of character recognition methods, and in particular relates to a general OCR training data generation system and method based on machine learning. Background technique [0002] With the development of machine learning and deep learning, in the field of optical character recognition (OCR), the deep learning algorithm is continuously iteratively updated, and the learning ability is continuously improved. Since deep learning is driven by data, a large amount of data is generated by combining data generation algorithms. , a good recognition effect can be obtained. [0003] However, the general OCR field involves many and complex scenes, mainly reflected in the fact that the placement angle of the text line in the image varies greatly in different scenes (horizontal, vertical, oblique and other angles); the background image is becoming more and more complex ( Complex backgrounds such as various patterns and color mixtures...

AI Technical Summary

Problems solved by technology

[0006] 1. The existing data generation algorithm is mainly to solve the training data generation of horizontal text, and does not generate vertical text and text data with a large oblique angle, resulting in such data (such as name plaques, billboards, etc.) The recognition effect is very poor in

[0007] 2. The background of the existing data generation algorithm is a specific scene picture background or a single solid color background, and the recognition effect of pictures under complex backgrounds including various patterns, mixed colors, etc. in general scene tasks is very poor

[0008] 3. The existing data generation algorithm fonts use specific fonts or specified fonts, and the image recognition effect of mixed fonts in data images (such as web page screenshots, billboards, etc.) is very poor

[0009] 4. When the picture and text are fused, the color difference algorithm is used to determine the text color and background color. Most of the current data contain complex backgrounds and texts of various colors, and the text color and background color have a good contrast. The color difference The algorithm is suitable for blending pictures with a single-color background. The contrast between the text and the background in the picture generated by the complex background is very poor, resulting in blurred text and poor recognition effect

Images