Character recognition method and device for image, electronic equipment and readable storage medium

Abstract

The invention provides a character recognition method and device for an image, electronic equipment and a readable storage medium, wherein the method comprises the steps: carrying out multi-scale detection of a to-be-processed image, to obtain a pixel-level thermodynamic diagram; carrying out instance segmentation based on the pixel-level thermodynamic diagram, and extracting textbox information corresponding to the to-be-processed image; according to each piece of textbox information, determining a textbox image corresponding to each piece of textbox information in the to-be-processed image;recognizing each textbox image, to obtain corresponding character recognition results. According to the invention, prediction and instance segmentation based on a pixel level are adopted; text of various angles and various curved shapes can be segmented in the to-be-processed image, interference of factors such as noise, wrinkles and non-uniform light in a complex scene is effectively resisted, then the whole text is recognized, recognition errors possibly caused by single-character segmentation are avoided, and the character recognition accuracy and recall rate of the image are remarkably improved.

Description

technical field [0001] The present application relates to the technical field of OCR (Optical Character Recognition, optical character recognition), specifically, the present application relates to an image character recognition method, device, electronic equipment and readable storage medium. Background technique [0002] OCR refers to the technology of identifying text information in an image. Its essence is to detect images captured by optical equipment and recognize text, extending vision and literacy capabilities to machines. OCR technology has been widely used in medical, insurance, finance, logistics, traditional manufacturing and other fields. For example, in an appointment scenario of the medical and health business, it is necessary to recognize the text in the clinical medical record photos taken and uploaded by the user with a mobile phone to achieve accurate appointments. The OCR technology can save considerable recognition and discrimination time, save a lot o...

AI Technical Summary

Problems solved by technology

[0003] However, in common business scenarios, the text information of photos is generally more complex, such as shadows, occlusions, wrinkles, distortions, multi-directional, multi-row dense, etc.

The ideal recognition effect cannot be achieved by using the image detection method in the prior art

For example, traditional object detection algorithms cannot adapt to the needs of text spanning the entire page

For another example, the image detection method of semantic segmentation only divides the overall area where the same type of target is located, which may cause relatively close multi-line texts to be unable to be effectively distinguished, resulting in the detection of multiple lines as one line, affecting subsequent identification process

Images