Unlock instant, AI-driven research and patent intelligence for your innovation.

Medical bill type text detection and recognition method and system based on deep neural network

A deep neural network and text detection technology, applied in the field of text detection and recognition of medical bills, can solve the problems of low accuracy, too small bills, influence of accuracy, etc., to facilitate the processing of feature extraction, improve efficiency and accuracy, The effect of ultra-high accuracy

Inactive Publication Date: 2018-11-30
深源恒际科技有限公司
View PDF5 Cites 33 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Under each major category, there are subcategories of different styles in each province and hospital. The number of subcategories exceeds hundreds, and the styles of bills of different subcategories may vary greatly. In addition, the background of the bill itself is relatively complicated, which leads to the traditional identification It is very difficult for the system to automatically recognize the bill text
Therefore, the traditional text detection and text recognition methods will have the problem of low accuracy rate; although for the text in a simple environment, the traditional method has a high recognition accuracy rate, but due to some characteristics of the bill itself, such as the shooting angle is non-horizontal, If the bill is too small, the accuracy of traditional identification methods will be greatly affected

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Medical bill type text detection and recognition method and system based on deep neural network
  • Medical bill type text detection and recognition method and system based on deep neural network
  • Medical bill type text detection and recognition method and system based on deep neural network

Examples

Experimental program
Comparison scheme
Effect test

no. 1 example

[0037] see figure 1 , figure 1 Shows the flowchart of the medical bill text detection and recognition method based on the deep neural network in this embodiment, the medical bill text detection and recognition method based on the deep neural network includes the following steps:

[0038] S101, using a target detection algorithm to detect the position of the bill in the bill picture;

[0039] S102, cropping the bill image according to the detected position of the bill;

[0040] S103, performing edge detection on the cropped receipt image, and rotating the receipt image to the horizontal direction according to the angles of the horizontal and vertical lines in the receipt image;

[0041] S104, performing text detection and recognition on the cropped and rotated bill image;

[0042] S105. Perform screening and extraction according to the recognized characters and orientations to obtain required data.

[0043] Further, in this embodiment, the target detection algorithm in S101...

no. 2 example

[0068] see image 3 , image 3 Shown is a structural block diagram of a medical bill text detection and recognition system 300 based on a deep neural network in this implementation, and the medical bill text detection and recognition system 300 based on a deep neural network includes the following structure:

[0069] The target detection module 301 is configured to use a target detection algorithm to detect the location of the bill in the bill picture;

[0070] A cropping module 302, configured to crop the bill picture according to the detected position of the bill;

[0071] The rotation module 303 is configured to perform edge detection on the cropped bill picture, and rotate the bill picture to the horizontal direction according to the angle of the horizontal and vertical straight lines in the bill picture;

[0072] Text detection and recognition module 304, used to perform text detection and recognition on the bill image after cutting and rotation;

[0073] The text scre...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a medical bill type text detection and recognition method and system based on a deep neural network. The method comprises the steps of, by using a target detection algorithm, detecting the position of a bill in a bill picture, and cutting the bill picture; performing edge detection on the cut bill picture, and rotating the bill picture to be in a horizontal direction according to horizontal and vertical straight line angles in the bill picture; and carrying out text detection and recognition on the cut and rotated bill picture; and performing screening and extracting according to recognized texts and directions so as to obtain required data. According to the method and the system, the methods of target detection, text recognition and the like are combined, so that the efficiency and accuracy of data entry can be improved. Moreover, based on the deep neural network, the extremely high accuracy can be obtained in text detection and text recognition; and the picture is subjected to processing convenient for feature extraction by using a traditional picture processing algorithm, so that the ultra-high accuracy also can be achieved in a complex scene of the medical bill.

Description

technical field [0001] The present invention relates to the technical field of text recognition, in particular to a method and system for detecting and recognizing medical bill texts based on a deep neural network. Background technique [0002] When an insurance company handles medical insurance cases, users will provide a large number of pictures of reimbursement receipts or copies, and the insurance company needs to enter the medical data required for the insurance case based on the content of these pictures. At present, the text entry of medical bills relies on manual methods to identify the contents of the bill pictures, and then realize the entry of medical data required for insurance cases. Therefore, this part of the work requires a lot of manpower, and the labor cost is extremely high. The entry process takes too long and is inefficient. [0003] In addition, the common medical bills of insurance companies are mainly divided into 4 categories: outpatient bills, hosp...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06K9/34G06K9/46G06K9/32G06N3/04
CPCG06V30/1478G06V30/153G06V10/267G06V10/44G06N3/045
Inventor 夏路遥
Owner 深源恒际科技有限公司