Granted Patent B2

US 10,089,555 · App. 14/229,108 · Granted Oct 2, 2018

Abstract

Claims (62)

1. A method executed by a processor of an optical character recognition (OCR) accuracy testing apparatus for testing an OCR process on a healthcare document, the method comprising:

receiving an original text file and a template for the healthcare document, the healthcare document being one of an electronic health record, a lab result, a patient visit record, a triage report, an emergency room report, or a surgery report, wherein the original text file comprises machine-readable text, wherein the template defines a font and a layout to be applied to the text in the original text file;

automatically converting the original text file according to the font and the layout defined by the template to generate a plurality of images of different file formats, each image in the plurality of images having pixels indicative of the text, an associated resolution, and the font and the layout defined by the template;

causing the OCR process to be executed for each image in the plurality of images to generate a converted text file for each image, the converted text file for each image comprising second machine-readable text, wherein the second machine-readable text is based upon the pixels;

calculating, with the processor, an accuracy score of the OCR process for each image in the plurality of the images based on a comparison of the text in the original text file and the second text in the converted text file for each image; and

flagging an image in the plurality of images when an accuracy score of the OCR process associated with the image is below a threshold score, wherein flagging the image indicates that a problem exists in the OCR process with respect to a file format of the image and the template.

2. The method of claim 1 , wherein converting the original text file comprises:

prior to generating the plurality of images of different file formats, converting the original text file to generate a plurality of intermediary word processing input files, wherein a font and a layout of an intermediary word processing input file in the plurality of intermediary word processing input files is based on the template; and

converting the intermediary word processing input file to the image.

3. The method of claim 1 , wherein the calculating the accuracy score of the OCR process comprises:

calculating a total number of words in the original text file;

calculating a total number of words in the converted text file;

comparing each word in the converted text file to the original text file to determine whether there is an exact match; and

calculating a word accuracy percentage based on the determinations of exact matches and based on the total number of words in the original text file and the converted text file.

4. The method of claim 1 , wherein a name of the healthcare document is included in a name of the original text file.

5. The method of claim 1 , wherein a name of the healthcare document is stored in association with an identifier of the original text file.

6. The method of claim 1 , wherein the layout is encoded with symbols or delimiters.

7. The method of claim 1 , wherein the layout defined by the template comprises predefined columns or tables.

8. The method of claim 1 , wherein the font and layout are applied to the text during the conversion of the original text file to the plurality of images of different file formats.

9. The method of claim 1 further comprising:

subsequent to automatically converting the original text file and prior to causing the OCR process to be executed for each image in the plurality of images, transmitting the plurality of images of different file formats to an OCR computing system that is in network communication with the OCR accuracy testing apparatus, wherein the OCR computing system executes the OCR process, wherein the OCR computing system transmits the converted text file for each image to the OCR accuracy testing apparatus; and

receiving the converted text file for each image from the OCR accuracy testing apparatus.

10. The method of claim 1 further comprising:

transmitting the accuracy score of the OCR process for each image in the plurality of images to a user terminal computing device operated by a user, wherein the user terminal computing device presents the accuracy of the OCR process for each image on a display.

11. A computer program product comprising at least one nontransitory computer-readable medium having computer-readable program instructions stored therein, the computer-readable program instructions comprising instructions, which when executed by a processor of an optical character recognition (OCR) accuracy testing apparatus, are configured to cause the OCR accuracy testing apparatus to test an OCR process on a healthcare document by performing acts comprising:

receiving an original text file and a template for the healthcare document, the healthcare document being one of an electronic health record, a lab result, a patient visit record, a triage report, an emergency room report, or a surgery report, wherein the original text file comprises machine-readable text, wherein the template defines at least a font and a layout to be applied to the text in the original text file;

automatically converting the original text file according to the font and the layout defined by the template to generate a plurality of images of different file formats, each image in the plurality of images having pixels indicative of the text, an associated resolution, and the font and the layout defined by the template;

causing the OCR process to be executed for each image in the plurality of images to generate a converted text file for each image, the converted text file for each image comprising second machine-readable text, wherein the second machine-readable text is based upon the pixels;

calculating, with the processor, an accuracy score of the OCR process for each image in the plurality of the images based on a comparison of the text in the original text file and the second text in the converted text file for each image; and

flagging an image in the plurality of images when an accuracy score of the OCR process associated with the image is below a threshold score, wherein flagging the image indicates that a problem exists in the OCR process with respect to a file format of the image and the template.

12. The computer program product of claim 11 , wherein converting the original text file comprises:

prior to generating the plurality of images of different file formats, converting the original text file to generate a plurality of intermediary word processing input files, wherein a font and a layout of an intermediary word processing input file in the plurality of intermediary word processing input files is based on the template; and

converting the intermediary word processing input file to the image.

13. The computer program product of claim 11 , wherein the calculating the accuracy score of the OCR process comprises:

calculating a total number of words in the original text file;

calculating a total number of words in the converted text file;

comparing each word in the converted text file to the original text file to determine whether there is an exact match; and

calculating a word accuracy percentage based on the determinations of exact matches and based on the total number of words in the original text file and the converted text file.

14. The computer program product of claim 11 , the acts further comprising:

subsequent to automatically converting the original text file and prior to causing the OCR process to be executed for each image in the plurality of images, transmitting the plurality of images of different file formats to an OCR computing system that is network communication with the OCR accuracy testing apparatus, wherein the OCR computing system executes the OCR process, wherein the OCR computing system transmits the converted text file for each image to the OCR accuracy testing apparatus; and

receiving the converted text file for each image from the OCR accuracy testing apparatus.

15. The computer program product of claim 11 , the acts further comprising:

transmitting the accuracy score of the OCR process for each image in the plurality of images to a user terminal computing device operated by a user, wherein the user terminal computing device presents the accuracy of the OCR process for each image on a display.

16. An optical character recognition (OCR) accuracy testing apparatus comprising processing circuitry comprising at least a processor configured to cause the OCR accuracy testing apparatus to test an OCR process on a healthcare document by performing acts comprising:

receiving an original text file and a template for the healthcare document, the healthcare document being one of an electronic health record, a lab result, a patient visit record, a triage report, an emergency room report, or a surgery report, wherein the original text file comprises machine-readable text, wherein the template defines at least a font and a layout to be applied to the text in the original text file;

converting the original text file according to the font and the layout defined by the template to generate a plurality of images of different file formats, each image in the plurality of images having pixels indicative of the text, an associated resolution, and the font and the layout defined by the template;

causing the OCR process to be executed for each image in the plurality of images to generate a converted text file for each image, the converted text file for each image comprising second machine-readable text, wherein the second machine-readable text is based upon the pixels;

calculating, with the processor, an accuracy score of the OCR process for each image in the plurality of the images based on a comparison of the text in the original text file and the second text in the converted text file for each image; and

flagging an image in the plurality of images when an accuracy score of the OCR process associated with the image is below a threshold score, wherein flagging the image indicates that a problem exists in the OCR process with respect to a file format of the image and the template.

17. The OCR accuracy testing apparatus of claim 16 , wherein converting the original text file comprises:

prior to generating the plurality of images of different file formats, converting the original text file to generate a plurality of intermediary word processing input files, wherein a font and a layout of an intermediary word processing input file in the plurality of intermediary word processing input files is based on the template; and

converting the intermediary word processing input file to the image.

18. The OCR accuracy testing apparatus of claim 16 , wherein the calculating the accuracy score of the OCR process comprises:

calculating a total number of words in the original text file;

calculating a total number of words in the converted text file;

comparing each word in the converted text file to the original text file to determine whether there is an exact match; and

calculating a word accuracy percentage based on the determinations of exact matches and based on the total number of words in the original text file and the converted text file.

19. The OCR accuracy testing apparatus of claim 16 , the acts further comprising:

subsequent to automatically converting the original text file and prior to causing the OCR process to be executed for each image in the plurality of images, transmitting the plurality of images of different file formats to an OCR computing system that is network communication with the OCR accuracy testing apparatus, wherein the OCR computing system executes the OCR process, wherein the OCR computing system transmits the converted text file for each image to the OCR accuracy testing apparatus; and

receiving the converted text file for each image from the OCR accuracy testing apparatus.

20. The OCR accuracy testing apparatus of claim 16 , the acts further comprising:

transmitting the accuracy score of the OCR process for each image in the plurality of images to a user terminal computing device operated by a user, wherein the user terminal computing device presents the accuracy of the OCR process for each image on a display.

Assignments (16)

Cited By (1)

US 12,567,211