Enhanced section detection using a combination of object detection with heuristics
Aspects of the invention include converting text from a first image file into a first machine-encodable text, wherein the image file includes a first section of text that is offset from a second section of text. Analyzing the first image file to detect a position of the first section of text. Embedding a first section of the first machine encodable-text with metadata describing the position of the first section of text. Reformatting the first section of the first machine encodable-text to conform to the position of the first section of text.
1. A computer-implemented method comprising:
detecting, by the processor, alphanumeric characters by recognizing patterns formed by dark areas of a first image file;
converting, by the processor, the detected alphanumeric characters to a first section of a machine-encodable text and a second section of a machine encodable-text;
detecting, by the processor, a position of the first section of the machine encodable-text, wherein detecting the position of the first section of the machine encodable-text comprises:
generating, by the processor, a plurality of bounding boxes in the first image file,
detecting, by the processor, features in a bounding box, of the plurality of bounding boxes, indicative of the first section of the machine encodable-text, and
classifying, by the processor, the bounding box as including the first section of the machine encodable-text;
detecting, by the processor, confidential information in the first section of the machine encodable-text;
replacing, by the processor, the confidential information with generic characters;
embedding, by the processor, the first section of the first machine encodable-text with metadata describing the position of the first section of text; and
reformatting, by the processor, the first section of the first machine encodable-text to conform to the position of the first section of text.
2. The computer-implemented method of claim 1 , wherein converting the text from the first image file to the first machine-encodable text comprises optical character recognition.
3. The computer-implemented method of claim 1 further comprising annotating the first section of machine-encodable text via a natural language processing technique.
4. The computer-implemented method of claim 1 , further comprising:
converting the first section of the first machine encodable-text to a second image file;
converting text of the second image file to a second machine-encodable text;
detecting a boundary of a subset of text from the second image file;
embedding the second machine encodable-text with metadata describing the boundary.
5. The computer-implemented method of claim 4 , wherein the subset comprises a sentence.
6. The computer implemented method of claim 4 , wherein detecting the boundary comprises performing a semantic analysis of text included in the second image file to determine the beginning and ending of the subset.
7. A system comprising:
a memory having computer readable instructions; and
one or more processors for executing the computer readable instructions, the computer readable instructions controlling the one or more processors to perform operations comprising:
detecting alphanumeric characters by recognizing patterns formed by dark areas of a first image file;
converting the detected alphanumeric characters to a first section of a machine-encodable text and a second section of a machine encodable-text;
detecting a position of the first section of the machine encodable-text, wherein detecting the position of the first section of the machine encodable-text comprises:
generating a plurality of bounding boxes in the first image file,
detecting features in a bounding box, of the plurality of bounding boxes, indicative of the first section of the machine encodable-text, and
classifying the bounding box as including the first section of the machine encodable-text;
detecting confidential information in the first section of the machine encodable-text;
replacing the confidential information with generic characters;
embedding the first section of the first machine encodable-text with metadata describing the position of first section; and
reformatting the first section of the machine encodable-text to conform to a position of the first section of text.
8. The system of claim 7 , wherein converting the text from the first image file to the first machine-encodable text comprises optical character recognition.
9. The system of claim 7 , wherein the operations further comprise annotating the first section of machine-encodable text via natural language processing techniques.
10. The system of claim 7 , wherein the operations further comprise:
converting the first section of the first machine encodable-text to a second image file;
converting text of the second image file to a second machine-encodable text;
detecting a boundary of a subset of text from the second image file;
embedding the second machine encodable-text with metadata describing the boundary.
11. The system of claim 10 , wherein the subset comprises a sentence.
12. The system of claim 10 , wherein detecting the boundary comprises performing a semantic analysis of text included in the second image file to determine the beginning and ending of the subset.
13. A computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to perform operations comprising:
detecting alphanumeric characters by recognizing patterns formed by dark areas of a first image file;
converting the detected alphanumeric characters to a first section of a machine-encodable text and a second section of a machine encodable-text;
detecting a position of the first section of the machine encodable-text, wherein detecting the position of the first section of the machine encodable-text comprises:
generating a plurality of bounding boxes in the first image file,
detecting features in a bounding box, of the plurality of bounding boxes, indicative of the first section of the machine encodable-text, and
classifying the bounding box as including the first section of the machine encodable-text;
detecting confidential information in the first section of the machine encodable-text;
replacing the confidential information with generic characters;
embedding the first section of the machine encodable-text with metadata describing the position of first section; and
reformatting the first section of the machine encodable-text to conform to a position of the first section of text.
14. The computer program product of claim 13 , wherein converting the text from the first image file to machine-encodable text comprises optical character recognition.
15. The computer program product of claim 13 , wherein the operations further comprise annotating the first section of machine-encodable text via natural language processing techniques.
16. The computer program product of claim 13 , wherein the operations further comprise:
converting the first section of the first machine encodable-text to a second image file;
converting text of the second image file to a machine-encodable text;
detecting a boundary of a subset of text from the second image file;
embedding the second machine encodable-text with metadata describing the boundary.
17. The computer program product of claim 16 , wherein the subset comprises a sentence.