Recognizing text in image data
A device may receive image data representing a document, the document including: text, and edges. Based on the edges, the device may identify, a segment of interest within the image data and crop the segment of interest to obtain a portion of the image data. In addition, the device may perform optical character recognition on the portion of the image data, the optical character recognition producing recognized text. The device may obtain, based on the recognized text, validation data that includes verification text, and determine whether the recognized text is verified based on the verification text. Based on a result of the determination, the device may perform an action.
1. A device, comprising:
one or more memory devices; and
one or more processors, communicatively connected to the one or more memory devices, to:
identify a plurality of edges within a document depicted in image data,
wherein the one or more processors, to identify the plurality of edges, are configured to:
determine artificial edges within the document, and
determine the plurality of edges based on the artificial edges;
identify, based on the plurality of edges, a segment of interest within the image data;
crop the segment of interest to obtain a portion of the image data and exclude edges, of the plurality of edges, that correspond to boxes or lines identified using one or more computer vision techniques to enable optical character recognition to be performed on the segment of interest unhindered by the boxes or the lines;
perform the optical character recognition on the portion of the image data via one or more optical character recognition models to generate a result based on the optical character recognition;
compare the result of the optical character recognition and validation data to verify the result of the optical character recognition based on using the result of the optical character recognition to identify the validation data; and
retrain, based on the image data and comparing the result of the optical character recognition and the validation data, the one or more optical character recognition models to recognize text included in another segment of interest that is similar to the segment of interest.
2. The device of claim 1 , wherein the plurality of edges correspond to one or more boxes or lines for text in a form associated with the image data.
3. The device of claim 1 , wherein the segment of interest is defined by a rectangular shape formed by the plurality of edges.
4. The device of claim 1 , wherein the image data depicts one or more of:
an image of a form,
an image of a word processing document,
a portable document format (PDF) image, or
a screen capture of a web page.
5. The device of claim 1 , wherein the one or more processors are further to:
discard portions of the image data that do not represent the segment of interest.
6. The device of claim 1 , wherein the one or more processors are further to:
obtain information identifying expected text within the segment of interest; and
provide the information identifying the expected text to the one or more optical character recognition models.
7. The device of claim 1 , wherein the one or more processors, when performing the optical character recognition, are to:
combine a plurality of respective results of the one or more optical character recognition models to determine the result of the optical character recognition.
8. A non-transitory computer-readable medium storing a set of instructions, the set of instructions comprising:
one or more instructions that, when executed by one or more processors of a device, cause the device to:
identify a plurality of edges within a document depicted in image data,
wherein the one or more instructions, that cause the one or more processors to identify the plurality of edges, cause the one or more processors to:
determine artificial edges within the document, and
determine the plurality of edges based on the artificial edges;
identify, based on the plurality of edges, a segment of interest within the image data;
crop the segment of interest to obtain a portion of the image data and exclude edges, of the plurality of edges, that correspond to boxes or lines identified using one or more computer vision techniques to enable optical character recognition to be performed on the segment of interest unhindered by the boxes or the lines;
perform the optical character recognition on the portion of the image data via one or more optical character recognition models to generate a result based on the optical character recognition;
compare the result of the optical character recognition and validation data to verify the result of the optical character recognition based on using the result of the optical character recognition to identify the validation data; and
retrain, based on the image data and comparing the result of the optical character recognition and the validation data, the one or more optical character recognition models to recognize text included in another segment of interest that is similar to the segment of interest.
9. The non-transitory computer-readable medium of claim 8 , wherein the plurality of edges correspond to one or more boxes or lines for text in a form associated with the image data.
10. The non-transitory computer-readable medium of claim 8 , wherein the segment of interest is defined by a rectangular shape formed by the plurality of edges.
11. The non-transitory computer-readable medium of claim 8 , wherein the image data depicts one or more of:
an image of a form,
an image of a word processing document,
a portable document format (PDF) image, or
a screen capture of a web page.
12. The non-transitory computer-readable medium of claim 8 , wherein the one or more instructions further cause the device to:
discard portions of the image data that do not represent the segment of interest.
13. The non-transitory computer-readable medium of claim 8 , wherein the one or more instructions further cause the device to:
obtain information identifying expected text within the segment of interest; and
provide the information identifying the expected text to the one or more optical character recognition models.
14. The non-transitory computer-readable medium of claim 8 , wherein the one or more instructions, that cause the device to perform the optical character recognition, cause the device to:
combine a plurality of respective results of the one or more optical character recognition models to determine the result of the optical character recognition.
15. A method, comprising:
identifying, by a device, a plurality of edges within a document depicted in image data,
wherein identifying the plurality of edges comprises:
determining artificial edges within the document, and
determining the plurality of edges based on the artificial edges;
identifying, by the device and based on the plurality of edges, a segment of interest within the image data;
cropping, by the device, the segment of interest to obtain a portion of the image data and exclude edges, of the plurality of edges, that correspond to boxes or lines identified using one or more computer vision techniques to enable optical character recognition to be performed on the segment of interest unhindered by the boxes or the lines;
performing, by the device, the optical character recognition on the portion of the image data via one or more optical character recognition models to generate a result based on the optical character recognition;
comparing, by the device, the result of the optical character recognition and validation data to verify the result of the optical character recognition based on using the result of the optical character recognition to identify the validation data; and
retraining, by the device and based on the image data and comparing the result of the optical character recognition and the validation data, the one or more optical character recognition models to recognize text included in another segment of interest that is similar to the segment of interest.
16. The method of claim 15 , wherein the plurality of edges correspond to one or more boxes or lines for text in a form associated with the image data.
17. The method of claim 15 , wherein the segment of interest is defined by a rectangular shape formed by the plurality of edges.
18. The method of claim 15 , wherein the image data depicts one or more of:
an image of a form,
an image of a word processing document,
a portable document format (PDF) image, or
a screen capture of a web page.
19. The method of claim 15 , further comprising:
discarding portions of the image data that do not represent the segment of interest.
20. The method of claim 15 , further comprising:
obtaining information identifying expected text within the segment of interest; and
providing the information identifying the expected text to the one or more optical character recognition models.