Method and system for optical character recognition of series of images

1. A method, comprising:

receiving, by a processing device, a current image of a series of images of an original document, wherein the current image at least partially overlaps with a previous image of the series of images;

performing optical character recognition (OCR) of the current image to produce an OCR text and a corresponding text layout;

associating, using a coordinate transformation, at least part of the OCR text with a first cluster of a plurality of clusters of symbol sequences, wherein the OCR text is produced by processing the current image and wherein the symbol sequences are produced by processing one or more previously received images of the series of images;

identifying a first median string representing the first cluster of symbol sequences based on a first subset of images of the series of images;

identifying a first template field of a document template corresponding to the first cluster based on the first median string representing the first cluster and the text layout of the current image;

analyzing the symbol sequences from the first cluster to identify suitable symbol sequences, wherein the suitable symbol sequences satisfy first parameters of the first template field;

identifying, for the first cluster, a second-level median string representing the cluster of symbol sequences based on a plurality of the suitable symbol sequences; and

producing, using the second-level median string, a resulting OCR text representing at least a portion of the first template field of the original document.

2. The method of claim 1 further comprising:

normalizing one or more symbol sequences from the first cluster to conform these symbol sequences to the first parameters of the first template field;

identifying, for the first cluster, the second-level median string representing the cluster of symbol sequences based on a plurality of the normalized symbol sequences.

3. The method of claim 2 wherein normalizing comprises:

identifying one or more unconforming symbols in the symbol sequence, wherein the unconforming symbol is a symbol that does not satisfy first parameters of the first template field;

replace the unconforming symbols in the symbol sequence with blank spaces.

4. The method of claim 1 wherein analyzing the symbol sequences from the first cluster to identify suitable symbol sequences comprises:

filtering the symbol sequences from the first cluster to filter out “garbage” symbol sequences.

5. The method of claim 1 , wherein the second-level median string is a constrained median string, satisfying second parameters of the first template field.

6. The method of claim 1 , further comprising:

determining that the second-level median string cannot be identified;

identifying, for the cluster, a third median string representing the cluster of symbol sequences based on a second subset of images of the series of images different from the first subset of images.

7. The method of claim 6 wherein the second subset of the series of images comprises the first subset of the series of images.

8. The method of claim 1 further comprising:

identifying the document template corresponding the original document based at least on the text layout, produced by the OCR, and the identified first median string.

9. The method of claim 8 wherein the document template is selected from a set of document templates.

10. The method of claim 8 further comprising:

determining that the document template for the original document cannot be identified with a satisfactory confidence level;

identifying a fourth median string representing the cluster of symbol sequences based on a third subset of images of the series of images different from the first subset of images.

11. A system, comprising:

a memory;

a processing device, coupled to the memory, the processing device configured to:

receive, by a processing device, a current image of a series of images of an original document, wherein the current image at least partially overlaps with a previous image of the series of images;

perform optical character recognition (OCR) of the current image to produce an OCR text and a corresponding text layout;

associate, using a coordinate transformation, at least part of the OCR text with a first cluster of a plurality of clusters of symbol sequences, wherein the OCR text is produced by processing the current image and wherein the symbol sequences are produced by processing one or more previously received images of the series of images;

identify a first median string representing the first cluster of symbol sequences based on a first subset of images of the series of images;

identify a first template field of a document template corresponding to the first cluster based on the first median string representing the first cluster and the text layout of the current image;

analyze the symbol sequences from the first cluster to identify suitable symbol sequences, wherein the suitable symbol sequences satisfy first parameters of the first template field;

identify, for the first cluster, a second-level median string representing the cluster of symbol sequences based on a plurality of the suitable symbol sequences; and

produce, using the second-level median string, a resulting OCR text representing at least a portion of the first template field of the original document.

12. The system of claim 11 further comprising:

normalizing one or more symbol sequences from the first cluster to conform these symbol sequences to the first parameters of the first template field;

identifying, for the first cluster, the second-level median string representing the cluster of symbol sequences based on a plurality of the normalized symbol sequences.

13. The system of claim 12 wherein normalizing comprises:

identifying one or more unconforming symbols in the symbol sequence, wherein the unconforming symbol is a symbol that does not satisfy first parameters of the first template field;

replace the unconforming symbols in the symbol sequence with blank spaces.

14. The system of claim 11 wherein analyzing the symbol sequences from the first cluster to identify suitable symbol sequences comprises:

filtering the symbol sequences from the first cluster to filter out “garbage” symbol sequences.

15. The system of claim 11 , wherein the second-level median string is a constrained median string, satisfying second parameters of the first template field.

16. The system of claim 11 , further comprising:

determining that the second-level median string cannot be identified;

identifying, for the cluster, a third median string representing the cluster of symbol sequences based on a second subset of images of the series of images different from the first subset of images.

17. The system of claim 11 further comprising:

identifying the document template corresponding the original document based at least on the text layout, produced by the OCR and the identified first median string.

18. The system of claim 17 further comprising:

determining that the document template for the original document cannot be identified with a satisfactory confidence level;

identifying a fourth median string representing the cluster of symbol sequences based on a third subset of images of the series of images different from the first subset of images.

19. A computer-readable non-transitory storage medium comprising executable instructions that, when executed by a processing device, cause the processing device to:

receive, by a processing device, a current image of a series of images of an original document, wherein the current image at least partially overlaps with a previous image of the series of images;

performed optical character recognition (OCR) of the current image to produce an OCR text and a corresponding text layout;

associate, using a coordinate transformation, at least part of the OCR text with a first cluster of a plurality of clusters of symbol sequences, wherein the OCR text is produced by processing the current image and wherein the symbol sequences are produced by processing one or more previously received images of the series of images;

identify a first median string representing the first cluster of symbol sequences based on a first subset of images of the series of images;

identify a first template field of a document template corresponding to the first cluster based on the first median string representing the first cluster and the text layout of the current image;

analyze the symbol sequences from the first cluster to identify suitable symbol sequences, wherein the suitable symbol sequences satisfy first parameters of the first template field;

identify, for the first cluster, a second-level median string representing the cluster of symbol sequences based on a plurality of the suitable symbol sequences; and

produce, using the second-level median string, a resulting OCR text representing at least a portion of the first template field of the original document.

20. The system of claim 19 further comprising:

normalizing one or more symbol sequences from the first cluster to conform these symbol sequences to the first parameters of the first template field;

identifying, for the first cluster, the second-level median string representing the cluster of symbol sequences based on a plurality of the normalized symbol sequences.