Machine learning based classification and annotation of paragraph of resume document images based on visual properties of the resume document images, and methods and apparatus for the same

1. A method, comprising:

generating a resume document image having a standardized format, based on a resume document having a plurality of paragraphs;

executing a statistical model to generate an annotated resume document image from the resume document image, the annotated resume document image indicating a bounding box and a paragraph type, for a paragraph from a plurality of paragraphs of the annotated resume document image, the statistical model being trained, before the executing, based on (1) a plurality of resume document images not including the resume document image and (2) a plurality of annotated resume document images not including the annotated resume document image;

identifying a block of text in the resume document corresponding to the paragraph of the annotated resume document image; and

extracting the block of text from the resume document and associating the paragraph type to the block of text.

2. The method of claim 1 , wherein the standardized format includes at least one of a common background color, a common font color, a common font ligature, a common font size, a common page margin, or a common page border.

3. The method of claim 1 , wherein the paragraph type includes a summary paragraph type, an experience paragraph type, an education paragraph type, a credential paragraph type, an award paragraph type, or a publication paragraph type.

4. The method of claim 1 , wherein the annotated resume document image further indicates a paragraph type confidence score, the method further includes associating the paragraph type confidence score to the block of text.

5. The method of claim 1 , wherein the resume document is a first resume document in a first language, the resume document image is a first resume document image, the annotated resume document image is a first annotated resume document image, the method further comprising:

generating a second resume document image having the standardized format and a plurality of paragraphs, based on a resume document in a second language; and

executing the statistical model to generate a second annotated resume document image from the second resume document image.

6. The method of claim 1 , wherein the standardized format includes a common background color, a common font color, a common font ligature, a common font size, a common page margin, and a common page border.

7. A non-transitory processor-readable medium storing code representing instructions to be executed by a processor of a first compute device, the code comprising code to cause the processor to:

receive, from a second compute device remote from the first compute device, a statistical model trained, at the second compute device, based on (1) a plurality of resume document images having a standardized format and (2) a plurality of annotated resume document images;

receive a resume document having a plurality of paragraphs;

generate a resume document image, not included within the plurality of resume document images, having the standardized format, based on the resume document;

execute the statistical model to generate an annotated resume document image from the resume document image, the annotated resume document image (1) indicating a bounding box and a paragraph type, for a paragraph from the plurality of paragraphs of the annotated resume document image, and (2) not included in the plurality of annotated resume document images;

identify a block of text in the resume document corresponding to the paragraph of the annotated resume document image; and

extract the block of text from the resume document and associate the paragraph type to the block of text.

8. The non-transitory processor-readable medium of claim 7 , wherein the standardized format includes at least one of a common background color, a common font ligature, a common font size, a common page margin, or a common page border.

9. The non-transitory processor-readable medium of claim 7 , wherein the standardized format includes at least one of a common background color, a common font color, a common font ligature, a common font size, a common page margin, or a common page border.

10. The non-transitory processor-readable medium of claim 7 , wherein the paragraph type includes at least one of a summary paragraph type, an experience paragraph type, an education paragraph type, a credential paragraph type, an award paragraph type, or a publication paragraph type.

11. The non-transitory processor-readable medium of claim 7 , wherein the annotated resume document image further indicates a paragraph type confidence score, the method further includes associating the paragraph type confidence score to the block of text.

12. The non-transitory processor-readable medium of claim 7 , wherein the resume document is a first resume document in a first language, the resume document image is a first resume document image, the annotated resume document image is a first annotated resume document image, the medium further comprising code to:

receive a second resume document in a second language, the second resume document having a plurality of paragraphs;

generate a second resume document image having the standardized format, based on the resume document; and

execute the statistical model to generate a second annotated resume document image from the second resume document image.

13. The non-transitory processor-readable medium of claim 7 , wherein the bounding box is a rectangle defined by two coordinates of two opposite corners of the rectangle.

14. A method, comprising:

receiving a plurality of resume documents, each resume document from the plurality of resume documents having a plurality of paragraphs;

generating a plurality of resume document images having a standardized format, based on the plurality of resume documents, each resume document image from the plurality of resume document images having a plurality of paragraphs;

annotating, for each resume document images from the plurality of document images, the plurality of paragraphs of the plurality of resume document images with a plurality of bounding boxes to generate a plurality of annotated resume document images, each bounding box from the plurality of bounding boxes indicating at least a cartesian coordinate and a paragraph type from a plurality of paragraph types;

training a statistical model based on the plurality of resume document images and the plurality of annotated resume document images;

receiving a resume document not included within the plurality of resume documents;

generating a resume document image having the standardized format, based on the resume document not included within the plurality of resume documents; and

executing, after training, the statistical model to generate an annotated resume document image based on the resume document image not included within the plurality of annotated resume document images, the annotated resume document image indicating a bounding box and a paragraph type, for each paragraph of the resume document image.

15. The method of claim 14 , further comprising:

identifying, for a paragraph of the annotated resume document image, a block of text corresponding to the paragraph of the annotated resume document image, in the resume document not included within the plurality of resume documents; and

extracting the block of text from the resume document not included within the plurality of resume documents and associating the paragraph type for the paragraph of the annotated resume document image to a paragraph type of the block of text.

16. The method of claim 14 , wherein the standardized format includes at least one a common background color, a common font color, a common font ligature, a common font size, a common page margin, or a common page border.

17. The method of claim 14 , wherein the plurality of paragraph types include at least two of a summary paragraph type, an experience paragraph type, an education paragraph type, a credential paragraph type, an award paragraph type, or a publication paragraph type.

18. The method of claim 14 , wherein the annotated resume document image further indicates a paragraph type confidence score.

19. The method of claim 14 , wherein the resume document not included within the plurality of resume documents is a first resume document in a first language, the resume document image is a first resume document image, the annotated resume document image is a first annotated resume document image, the method further comprising:

receiving a second resume document in a second language, the second resume document having a plurality of paragraphs;

generating a second resume document image having the standardized format, based on the second resume document; and

executing, after training, the statistical model to generate a second annotated resume document image based on the second resume document image.

20. The method of claim 14 , wherein the bounding box is a rectangle defined by two coordinates of two opposite corners of the rectangle.