Text detection method of document image in natural scene

Abstract

The invention discloses a text detection method of a document image in a natural scene. Commonly used Chinese characters are selected to make Chinese character pictures, a dataset 1 is formed, randomrevolving and cropping operations are carried out on labeled document images, then a manner of Poisson cloning is used to fuse different background images, and a dataset 2 is formed; the dataset 1 isadopted to carry out training of a text classification model on a VGG16 network, and after the model converges, obtained parameters are used to initialize a fully convolutional neural network model, and the dataset 2 is used to train the model; the trained fully-convolutional neural network model is used to process the image, a classification situation of each pixel point is obtained according toa maximum probability method, and a text-non-text binary image is formed; a method of connected regions is used to obtain text regions, the original image is binarized, and only text information in the text regions in the text-non-text region binary image is extracted to obtain a text binary image; the image is corrected through a maximum variance method; and projection is carried out again on thecorrected image, and the text-non-text region binary image is refined.

Description

technical field [0001] The invention belongs to an image processing method, in particular to a text detection method of a certificate image in a natural scene. Background technique [0002] The rapid development of Internet technology and the popularization of smart phones have greatly facilitated our lives. In many scenarios, operators need users to upload documents (such as ID cards, business licenses, etc.) to verify the user's identity and qualifications. The user's mobile phone takes a picture of the certificate and uploads it for verification, which is convenient and efficient. Due to the complex background and many environmental interference factors of users taking pictures in natural scenes. There are various shooting backgrounds in natural scenes. Users may shoot in possible life scenes such as desktops and bed sheets with complex textures. These textures are difficult to distinguish from text. There are also cases where the text is partially occluded in the phot...

AI Technical Summary

Problems solved by technology

[0009] (1) The sliding window needs to traverse the image at multiple scales and judge each detection window, resulting in long detection time and low efficiency

[0010] (2) Lack of precision, difficult to deal with complex backgrounds

This method fails when the background is complex and the edge cannot be detected

At the same time, this method will mistakenly detect many objects with regular lines and similar to text, such as strips, grids, bricks, etc., as text

Unable to cope with the user's demand for shooting multiple scenes in a natural environment

[0012] MSER, does not work well with blurry, lighting, color, texture changes, low-contrast text

[0013] Both SWT and MSER methods detect a single character, and the detection result is inconvenient for the OCR module. It is necessary to merge the detected single characters according to the character spacing, height difference and other characteristics, which increases the amount of calculation.

Images