Feature classification with spatial analysis

1. A method for identifying features in a digital image, the method comprising:

accessing a set of spatial templates, each spatial template including a set of features and a set of bitmaps, each bitmap indicating a spatial relationship between each feature and respective remaining features in each spatial template;

in response to receiving the digital image to be processed, selecting a spatial template from the set of spatial templates;

segmenting the digital image into image segments;

extracting features from each image segment;

classifying each of the extracted features as one of the set of features in the spatial template;

in response to at least one feature being unclassified, generating at least one bitmap representing a spatial relationship for each of the at least one unclassified feature in relation to classified features;

comparing the at least one bitmap with each of a set of bitmaps in the selected spatial template; and

identifying the at least one unclassified feature as one of the set of features in the selected spatial template based on a result of the comparison.

2. The method of claim 1 , wherein the set of features includes at least one of a label, field, marker or ancillary text.

3. The method of claim 1 , wherein each bitmap comprises a combination of multi-bits, each multi-bit corresponding to a point of a compass defined between one feature and another feature.

4. The method of claim 1 , wherein each bitmap comprises a combination of multi-bits, each multi-bit corresponding to a vector defined between one feature and another feature, wherein a first portion of a multi-bit indicates a magnitude component of the vector, and a second portion of the multi-bit indicates a direction component of the vector.

5. The method of claim 1 , wherein comparing the at least one bitmap with each of a set of bitmaps comprises performing not-exclusive-or (NXOR) operations between the at least one bitmap and each of the set of bitmaps in the selected spatial template.

6. The method of claim 5 , wherein comparing the at least one bitmap with each of a set of bitmaps further comprises:

classifying the at least one unclassified feature as a certain feature in the spatial template when an NXOR operation with the certain feature produces a resultant bitmap with a maximum number of ‘1’ bits.

7. The method of claim 1 , further comprising creating the set of spatial templates and storing the set of spatial templates in the storage.

8. A non-transitory computer-readable storage medium containing instructions that, when executed by one or more processors, perform operations for classifying features in a digital image, the operations comprising:

accessing a storage storing a set of spatial templates, each spatial template including a set of features and a set of bitmaps, each bitmap indicating a spatial relationship between each feature and respective remaining features in each spatial template;

in response to receiving the digital image to be processed, selecting a spatial template from the set of spatial templates;

segmenting the digital image into image segments;

extracting features from each image segment;

classifying each of the extracted features as one of the set of features in the spatial template;

in response to at least one feature being unclassified, generating at least one bitmap representing a spatial relationship for each of the at least one unclassified feature in relation to classified features;

comparing the at least one bitmap with each of a set of bitmaps in the selected spatial template; and

identifying the at least one unclassified feature as one of the set of features in the selected spatial template based on a result of the comparison.

9. The non-transitory computer-readable storage medium of claim 8 , wherein the set of features includes at least one of a label, field, marker or ancillary text.

10. The non-transitory computer-readable storage medium of claim 8 , wherein each bitmap comprises a combination of multi-bits, each multi-bit corresponding to a point of a compass defined between one feature and another feature.

11. The non-transitory computer-readable storage medium of claim 8 , wherein each bitmap comprises a combination of multi-bits, each multi-bit corresponds to a vector defined between one feature and another feature, wherein a first portion of a multi-bit indicates a magnitude component of the vector, and a second portion of the multi-bit indicates a direction component of the vector.

12. The non-transitory computer-readable storage medium of claim 8 , wherein comparing the at least one bitmap with each of a set of bitmaps comprises performing not-exclusive-or (NXOR) operations between the at least one bitmap and each of the set of bitmaps in the selected spatial template.

13. The non-transitory computer-readable storage medium of claim 12 , wherein comparing the at least one bitmap with each of a set of bitmaps further comprises:

classifying the at least one unclassified feature as a certain feature in the spatial template when an NXOR operation with the certain feature produces a resultant bitmap with a maximum number of ‘1’ bits.

14. A system, comprising:

one or more processors; and

a memory storing one or more applications, which, when executed on the one or more processors perform operations for identifying form fields in a digital image, the operations comprising:

classifying features in a digital image, the operations comprising:

creating a set of spatial templates, each spatial template including a set of features and a set of bitmaps, each bitmap indicating a spatial relationship between each feature and respective remaining features in each spatial template;

in response to receiving the digital image to be processed, selecting a spatial template from the set of spatial templates;

segmenting the digital image into image segments;

extracting features from each image segment;

classifying each of the extracted features as one of the set of features in the spatial template;

in response to at least one feature being unclassified, generating at least one bitmap representing a spatial relationship for each of the at least one unclassified feature in relation to classified features;

comparing the at least one bitmap with each of a set of bitmaps in the selected spatial template; and

identifying the at least one unclassified feature as one of the set of features in the selected spatial template based on a result of the comparison.

15. The system of claim 14 , wherein the set of features includes at least one of a label, field, marker or ancillary text.

16. The system of claim 14 , wherein each bitmap comprises a combination of multi-bits, each multi-bit corresponding to a point of a compass defined between one feature and another feature.

17. The system of claim 14 , wherein each bitmap comprises a combination of multi-bits, each multi-bit corresponding to a vector defined between one feature and another feature, wherein a first portion of a multi-bit indicates a magnitude component of the vector, and a second portion of the multi-bit indicates a direction component of the vector.

18. The system of claim 14 , wherein comparing the at least one bitmap with each of a set of bitmaps comprises performing not-exclusive-or (NXOR) operations between the at least one bitmap and each of the set of bitmaps in the selected spatial template.

19. The system of claim 18 , wherein comparing the at least one bitmap with each of a set of bitmaps further comprises:

classifying the at least one unclassified feature as a certain feature in the spatial template, in which an NXOR operation with the certain feature produces a resultant bitmap with a maximum number of ‘1’ bits.

20. The system of claim 14 , further comprising a storage configured to store the set of spatial templates.