Character detection method and system

Abstract

The invention discloses a character detection method and system, and the method comprises the steps: carrying out the feature extraction of an input image, and obtaining a feature image; predicting byusing the adaptive regional suggestion network to obtain a suggestion box; utilizing the suggestion box to cut the feature image to obtain a cut feature map; respectively carrying out character texture information modeling on the cutting feature map in two orthogonal directions to obtain a contour point thermodynamic diagram corresponding to each orthogonal direction; and screening contour pointsin the contour point thermodynamic diagram to obtain a contour point set so as to reconstruct characters in the input image. The adaptive region suggestion network can adapt to the scale change of characters to generate a suggestion box corresponding to a character region, and the character texture information modeling module performs character texture information modeling in the orthogonal direction to suppress false positive contour points, so that the detection precision of characters in any shape scene is improved.

Description

technical field [0001] The present disclosure relates to the technical field of text recognition, and in particular, to a text detection method and system. Background technique [0002] Text detection in natural scenes refers to the detection of text areas in complex backgrounds and marking the text areas with bounding boxes. The results of text detection in natural scenes are widely used in autonomous driving, robotics and other fields. Text detection in natural scenes faces difficulties such as low resolution, complex background, and variable font size, which makes the practical application effect of traditional text detection technology poor. [0003] With the development of deep learning technology, the natural scene text detection technology based on deep learning has been significantly improved. Although this detection technology can detect text of any shape, the detection results contain many false positive detections and are affected by various text sizes. Due to t...

AI Technical Summary

Problems solved by technology

Text detection in natural scenes faces difficulties such as low resolution, complex background, and variable font size, which makes the actual application effect of traditional text detection technology poor

[0003] With the development of deep learning technology, the natural scene text detection technology based on deep learning has been significantly improved. Although this detection technology can detect text of any shape, the detection results contain many false positive detections and are affected by various text sizes. Influenced by sexual problems, its detection accuracy needs to be improved

Images