Text detection method and device, storage medium and computer equipment

Abstract

The invention provides a text detection method. The method comprises that an image to be detected is obtained; the image to be detected is input to a neural network model, and a target characteristicmatrix is output; the target characteristic matrix is input to a full connection layer, and the full connection layer map different elements of the target characteristic matrix into predicted image subareas corresponding to the image to be detected according to a preset anchor point area; and text characteristic information of the predicted image subareas is obtained, and the predicted image subareas are connected into a corresponding predicted text row via a text clustering algorithm according to the text characteristic information of the predicted image subareas. The text area of the image to be detected is determined by using the predicted image subareas, and the detection accuracy and robustness are improved. The invention also provides a text detection device, a storage medium and computer equipment.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a text detection method, device, storage medium and computer equipment. Background technique [0002] In order to meet certain application requirements, it is necessary to detect the target object in the image. Since the usual object has a complete closed boundary, the usual target object detection method directly predicts the candidate area where the target object is located, and can be based on the predicted candidate area. Part of the characteristics of the target object infers the object category, and realizes the detection of the target object. [0003] However, text is different from ordinary objects. The boundary of text changes with strokes, and there may be spaces between a text. It is difficult to determine the type of text through a part of text. Therefore, it is easy to use traditional target object detection algorithms due to text Spaces cause false detection and ...

AI Technical Summary

Problems solved by technology

[0003] However, text is different from ordinary objects. The boundary of text changes with strokes, and there may be spaces between a text. It is difficult to determine the type of text through a part of text. Therefore, it is easy to use traditional target object detection algorithms due to text Spaces cause false detection and missed detection, and because the category of the entire text cannot be predicted based on some text in the prediction candidate area, the accuracy of text positioning is not high, and the detection robustness is low

Images