Model-based method of document logical structure recognition in OCR systems
View Patent ↗In one embodiment, the invention provides a method for determining a logical structure of a document. The method comprises generating at least one document hypothesis for the whole document; for each document hypothesis, verifying said document hypothesis including (a) generating at least one block hypothesis for each block in the document based on the document hypothesis; and (b) selecting a best block hypothesis for each block; selecting as a best document hypothesis the document hypothesis that has the best degree of correspondence with the selected best block hypotheses for the document; and forming the document based on the best document hypothesis.
1. A method for determining a logical structure of a document, the method comprising:
acquiring an image of the document;
identifying one or more blocks in the image of the document;
generating a hypothesis for at least one of the identified blocks in the image of the document (a “block hypothesis”);
generating at least one document hypothesis for the image of the document, wherein said generating included referencing a plurality of document models, wherein each document model describes one or more possible logical structures, and wherein such logical structures are based on the presence of one or more blocks;
selecting a document hypothesis based on its degree of correspondence with at least one block hypothesis; and
forming a representation of the document based on the selected document hypothesis.
2. The method of claim 1 , wherein the generating the at least one document hypothesis for the image of the document includes generating a plurality of document hypotheses in order of differing probabilities.
3. The method of claim 1 ,
wherein indentifying the one or more blocks includes performing a physical structure analysis to identify each said one or more blocks.
4. The method of claim 1 , the method further comprising:
saving the representation of the document in an extended format in a memory.
5. The method of claim 1 , wherein the generating the at least one document hypothesis for the image of the document is based on information about a possible arrangement of blocks in the image of the document.
6. The method of claim 1 , wherein said generating the hypothesis for at least one of the identified blocks in the document is based on the at least one document hypothesis.
7. The method of claim 1 , wherein generating the hypothesis for at least one of the identified blocks includes generating a hypothesis for each identified block in the image of the document.
8. The method of claim 1 , wherein a block comprises one or more form elements.
9. A system comprising:
a processor; and
a memory coupled to the processor, the memory storing instructions which when executed by the processor, cause the system to perform a method for determining a logical structure for a document, comprising:
identifying one or more blocks in an image of the document;
generating a hypothesis for at least one of the identified blocks in the image of the document;
generating at least one document hypothesis for the image of the document, wherein said generating includes referencing a plurality of document models, wherein each document model describes one or more possible logical structures, and wherein such logical structures are based on the presence of one or more blocks;
selecting a document hypothesis based on its degree of correspondence with at least one block hypothesis; and
forming a representation of the document based on the selected document hypothesis.
10. The system of claim 1 , wherein the generating the at least one document hypothesis for the document includes generating a plurality of document hypotheses in order of differing probabilities.
11. The system of claim 1 , wherein wherein identifying the one or more blocks includes performing a physical structure analysis to identify each said one or more blocks.
12. The system of claim 1 , wherein the method further comprises saving the representation of the document in an extended format in a memory.
13. The system of claim 1 , wherein the document hypothesis is selected automatically.
14. The system of claim 1 , wherein the document hypothesis is selected by receiving a selection through a user interface element.
15. The system of claim 1 , wherein the generating the at least one document hypothesis for the image of the document is based on information about a possible arrangement of blocks in the image of the document.
16. A non-transitory computer-readable medium having stored thereon instructions which when executed by a computer, cause the computer to perform a method for determining a logical structure for a document, comprising:
identifying one or more blocks in an image of the document;
generating a hypothesis for at least one of the identified blocks in the image of the document;
generating at least one document hypothesis for the image of the document, wherein said generating includes referencing a plurality of document models, wherein each document model describes one or more possible logical structures, and wherein such logical structures are based on the presence of one or more blocks;
selecting a document hypothesis based on its degree of correspondence with at least one block hypothesis; and
forming a representation of the document based on the selected document hypothesis.
17. The non-transitory computer-readable medium of claim 16 , wherein the generating includes generating a plurality of document hypotheses in order of differing probabilities.
18. The non-transitory computer-readable medium of claim 16 , wherein wherein identifying the one or more blocks includes performing a physical structure analysis to identify each said one or more blocks.
19. The non-transitory computer-readable medium of claim 16 , wherein the method further comprises saving a representation of the document in an extended format in a memory.
20. The non-transitory computer-readable medium of claim 16 , wherein the document hypothesis is selected automatically.
21. The non-transitory computer-readable medium of claim 16 , wherein the document hypothesis is selected by receiving a selection through a user interface element.
22. The method of claim 1 , wherein generating said at least one document hypothesis includes comparing the hypothesis for said at least one of the identified blocks against a block model representing a possible logical structure for a corresponding block.
23. The method of claim 1 , wherein generating said at least one document hypothesis includes comparing the hypothesis for said at least one of the identified blocks on the basis of a degree of correspondence between the hypothesis for said at least one of the identified blocks and each of a plurality of block models.