Automated systems and methods for identifying fields and regions of interest within a document image
Systems and methods are configured to extract text from images (e.g., document images) utilizing a combination of optical character recognition processes and neural network-based analysis of various images to train a machine-learning object recognition model that is configured to identify text within images based on object-comparisons between known and unknown text within images. In certain embodiments, identified text within images may be utilized to identify corresponding regions-of-interest for extraction of image data encompassing a portion of an image that may be indexed based at least in part on text identified as corresponding to the particular region-of-interest.
1. A computer-implemented method for extracting image data within a region-of-interest from an image, the method comprising:
identifying a plurality of fields within an image;
extracting text from a first field of the plurality of fields;
identifying at least one neighbor field;
determining, within the image, a location of the first field and a location of the at least one neighbor field;
identifying boundaries of a closed perimeter of the region-of-interest within the image, wherein:
a first boundary of the closed perimeter of the region-of-interest is aligned with a boundary of the first field; and
a second boundary of the closed perimeter of the region-of-interest is aligned with a boundary of one of the at least one neighbor field;
extracting, from the image, image data of a portion of the image bounded by the closed perimeter of the region-of-interest within the image; and
storing the extracted image data within a database.
2. The computer-implemented method of claim 1 , further comprising:
extracting text from each of the plurality of fields identified within the image; and
comparing the text extracted from each of the plurality of fields with one or more field identifiers to identify the first field of the plurality of fields.
3. The computer-implemented method of claim 1 , wherein extracting text from the first field comprises executing an optical character recognition (OCR) process on the first field.
4. The computer-implemented method of claim 1 , wherein identifying at least one neighbor field comprises:
determining location data for each of the plurality of fields; and
identifying a second field of the plurality of fields that is nearest to the first field in a horizontal direction within the image, wherein the second field is identified as a neighbor field of the at least one neighbor field.
5. The computer-implemented method of claim 1 , wherein identifying at least one neighbor field comprises:
determining location data for each of the plurality of fields; and
identifying a second field of the plurality of fields that is nearest to the first field in a vertical direction within the image, wherein the second field is identified as a neighbor field of the at least one neighbor field.
6. The computer-implemented method of claim 5 , wherein identifying at least one neighbor field further comprises:
identifying a third field of the plurality of fields that is nearest to the first field in a horizontal direction within the image, wherein the third field is identified as a neighbor field of the at least one neighbor field.
7. The computer-implemented method of claim 6 , wherein identifying boundaries of the closed perimeter of the region-of-interest within the image comprises:
identifying the first boundary of the closed perimeter of the region-of-interest as aligned with a first boundary of the first field;
identifying the second boundary of the closed perimeter of the region-of-interest as aligned with a boundary of the second field; and
identifying a third boundary of the closed perimeter of the region of interest as aligned with a boundary of the third field; and
identifying a fourth boundary of the closed perimeter of the region of interest as aligned with a second boundary of the first field; and
wherein the region-of-interest is rectangular.
8. The computer-implemented method of claim 1 , wherein identifying boundaries of the closed perimeter of the region-of-interest within the image comprises:
identifying a third boundary of the closed perimeter of the region of interest as aligned with an edge of the image.
9. A computing system comprising a non-transitory computer readable storage medium and one or more processors, the computing system configured to:
identify a plurality of fields within an image;
extract text from a first field of the plurality of fields;
identify at least one neighboring field;
determine, within the image, a location of the first field and a location of the at least one neighboring field;
identify boundaries of a closed perimeter of the region-of-interest within the image, wherein:
a first boundary of the closed perimeter of the region-of-interest is aligned with a boundary of the first field; and
a second boundary of the closed perimeter of the region-of-interest is aligned with a boundary of one of the at least one neighbor field;
extract, from the image, image data of a portion of the image bounded by the closed perimeter of the region-of-interest within the image; and
store the extracted image data within a database.
10. The computing system of claim 9 , wherein the computing system is further configured to:
extract text from each of the plurality of fields identified within the image; and
compare the text extracted from each of the plurality of fields with one or more field identifiers to identify the first field of the plurality of fields.
11. The computing system of claim 9 , wherein extracting text from the first field comprises executing an optical character recognition (OCR) process on the first field.
12. The computing system of claim 9 , wherein identifying at least one neighbor field comprises:
determining location data for each of the plurality of fields; and
identifying a second field of the plurality of fields that is nearest to the first field in a horizontal direction within the image, wherein the second field is identified as a neighbor field of the at least one neighbor field.
13. The computing system of claim 9 , wherein identifying at least one neighbor field comprises:
determining location data for each of the plurality of fields; and
identifying a second field of the plurality of fields that is nearest to the first field in a vertical direction within the image, wherein the second field is identified as a neighbor field of the at least one neighbor field.
14. The computing system of claim 13 , wherein identifying at least one neighbor field further comprises:
identifying a third field of the plurality of fields that is nearest to the first field in a horizontal direction within the image, wherein the third field is identified as a neighbor field of the at least one neighbor field.
15. The computing system of claim 14 , wherein identifying boundaries of the closed perimeter of the region-of-interest within the image comprises:
identifying the first boundary of the closed perimeter of the region-of-interest as aligned with a first boundary of the first field;
identifying the second boundary of the closed perimeter of the region-of-interest as aligned with a boundary of the second field; and
identifying a third boundary of the closed perimeter of the region of interest as aligned with a boundary of the third field; and
identifying a fourth boundary of the closed perimeter of the region of interest as aligned with a second boundary of the first field; and
wherein the region-of-interest is rectangular.
16. The computing system of claim 9 , wherein identifying boundaries of the closed perimeter of the region-of-interest within the image comprises:
identifying a third boundary of the closed perimeter of the region of interest as aligned with an edge of the image.
17. A computer program product comprising a non-transitory computer readable medium having computer program instructions stored therein, the computer program instructions when executed by a processor, cause the processor to:
identify a plurality of fields within an image;
extract text from a first field of the plurality of fields;
identify at least one neighboring field;
determine, within the image, a location of the first field and a location of the at least one neighboring field;
identify boundaries of a closed perimeter of the region-of-interest within the image, wherein:
a first boundary of the closed perimeter of the region-of-interest is aligned with a boundary of the first field; and
a second boundary of the closed perimeter of the region-of-interest is aligned with a boundary of one of the at least one neighbor field;
extract, from the image, image data of a portion of the image bounded by the closed perimeter of the region-of-interest within the image; and
store the extracted image data within a database.
18. The computer program product of claim 17 , wherein the computing system is further configured to:
extract text from each of the plurality of fields identified within the image; and
compare the text extracted from each of the plurality of fields with one or more field identifiers to identify the first field of the plurality of fields.
19. The computer program product of claim 17 , wherein extracting text from the first field comprises executing an optical character recognition (OCR) process on the first field.
20. The computer program product of claim 17 , wherein identifying at least one neighbor field comprises:
determining location data for each of the plurality of fields; and
identifying a second field of the plurality of fields that is nearest to the first field in a horizontal direction within the image, wherein the second field is identified as a neighbor field of the at least one neighbor field.
21. The computer program product of claim 17 , wherein identifying at least one neighbor field comprises:
determining location data for each of the plurality of fields; and
identifying a second field of the plurality of fields that is nearest to the first field in a vertical direction within the image, wherein the second field is identified as a neighbor field of the at least one neighbor field.
22. The computer program product of claim 21 , wherein identifying at least one neighbor field further comprises:
identifying a third field of the plurality of fields that is nearest to the first field in a horizontal direction within the image, wherein the third field is identified as a neighbor field of the at least one neighbor field.
23. The computer program product of claim 22 , wherein identifying boundaries of the closed perimeter of the region-of-interest within the image comprises:
identifying the first boundary of the closed perimeter of the region-of-interest as aligned with a first boundary of the first field;
identifying the second boundary of the closed perimeter of the region-of-interest as aligned with a boundary of the second field; and
identifying a third boundary of the closed perimeter of the region of interest as aligned with a boundary of the third field; and
identifying a fourth boundary of the closed perimeter of the region of interest as aligned with a second boundary of the first field; and
wherein the region-of-interest is rectangular.
24. The computer program product of claim 17 , wherein identifying boundaries of the closed perimeter of the region-of-interest within the image comprises:
identifying a third boundary of the closed perimeter of the region of interest as aligned with an edge of the image.