Method and device for determining text line region

Abstract

The invention discloses a method and a device for determining a text line region, and relates to the technical field of computers. According to one specific embodiment, the method comprises the steps of obtaining a text image, wherein a text in the text image is presented in a text line mode; detecting to obtain center line information of text lines by using a text line detection model based on center lines; and obtaining a text line contour according to the center line information so as to determine a text line region. According to the embodiment, the problem of text adhesion can be effectively solved, text lines can be accurately segmented by extracting center lines of the text lines especially under the condition that the text lines are very close to one another or have overlapped regions, the text line segmentation precision can be improved, and then the accuracy of subsequent text recognition is improved; moreover, the center lines of the text lines are determined firstly, then the text line contour (region) is obtained through detection according to the center line information, and the method is particularly suitable for text detection of long text lines.

Description

technical field [0001] The present invention relates to the field of computer technology, in particular to a method and device for determining text line regions. Background technique [0002] Text detection technology is an important part of OCR (Optical Character Recognition) technology, which mainly uses computer vision technology to detect text areas in images for subsequent operations such as text recognition. The current methods of text detection can be roughly divided into three types: methods based on candidate frame regression, methods based on image segmentation, and methods combined with segmentation and regression. [0003] In the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art: The existing segmentation-based methods mainly segment the text confidence map, although the shrink (shrink) strategy can avoid text cohesion as much as possible, However, when the text lines are very close...

AI Technical Summary

Problems solved by technology

[0003] In the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art: The existing segmentation-based methods mainly segment the text confidence map, although the shrink (shrink) strategy can avoid text cohesion as much as possible, However, when the text lines are very close or there are overlapping areas between the text lines, the existing technical solutions usually cannot segment the text lines well, resulting in text detection errors, which in turn affects the accuracy of subsequent text recognition.

Images