Column inferencer based on generated border pieces and column borders
A method for processing an electronic document (ED) to infer columns in the ED, where the ED comprises a plurality of characters. The method includes generating a mark-up version of the ED having text-layout attributes of the characters in the ED, where the characters are grouped into paragraphs based on the text-layout attributes, and each paragraph corresponds to a paragraph bounding box surrounding a corresponding paragraph, generating border pieces by initiating a pair of left scan and right scan from each paragraph bounding box to identify any adjacent paragraph bounding box, and generating, based at least on the border pieces, column borders for use in inferring the columns in the ED, where at least one column has a vertically aligned portion of the paragraphs.
1. A method for processing an electronic document (ED) to infer columns in the ED, wherein the ED comprises a plurality of characters, the method comprising:
generating a mark-up version of the ED comprising text-layout attributes of the characters in the ED, wherein
the characters are grouped into a plurality of paragraphs based on the text-layout attributes, and
each of the plurality of paragraphs corresponds to a paragraph bounding box surrounding a corresponding paragraph;
generating a plurality of border pieces by initiating a pair of left scan and right scan from each of the plurality of paragraph bounding boxes to identify any adjacent paragraph bounding box, at least one scan of the left scan and the right scan being initiated from said first one of the two adjacent paragraph bounding boxes to reach a second one of the two adjacent paragraph bounding boxes, wherein at least one border piece of the plurality of border pieces comprises:
a white space separating two adjacent paragraph bounding boxes of the plurality of paragraph bounding boxes in the horizontal direction;
a horizontal dimension equaling a separation between the two adjacent paragraph bounding boxes;
a vertical dimension equaling a height of a first one of the two adjacent paragraph bounding boxes;
a direction of said at least one scan of the left scan and the right scan; and
generating, based at least on the plurality of border pieces, a plurality of column borders for use in inferring the columns in the ED,
wherein at least one column comprises a vertically aligned portion of the plurality of paragraphs.
2. The method of claim 1 , wherein generating the plurality of column borders comprises:
generating a sorted list of the plurality of border pieces based on respective locations of the plurality of border pieces;
generating a plurality of potential column borders by initiating a pair of forward traversal and reverse traversal of the sorted list from each of the plurality of border pieces, wherein the pair of forward traversal and reverse traversal identifies any overlapping border pieces for combining into one of the plurality of potential column borders; and
converting the plurality of potential column borders into the plurality of column borders.
3. The method of claim 2 ,
wherein each of the plurality of border pieces corresponds to one of the plurality of paragraphs,
wherein each of the plurality of potential column borders is associated with a list of traversed border pieces based on a corresponding pair of forward traversal and reverse traversal, and
wherein converting the plurality of potential column borders comprises:
removing duplication among the plurality of potential column borders based at least on the list of traversed border pieces of each of the plurality of potential column borders.
4. The method of claim 3 , further comprising:
inferring a column based on corresponding paragraphs of border pieces contained in one of the plurality of column borders.
5. The method of claim 2 , wherein converting the plurality of potential column borders comprises:
combining two or more of the plurality of potential column borders based on at least one selected from a group consisting of a horizontal overlap criterion and a vertical overlap criterion.
6. The method of claim 2 , wherein converting the plurality of potential column borders comprises:
generating, from the plurality of potential column borders, a column border group comprising two or more potential column borders based on a vertical overlap criterion,
wherein the column border group identifies a region of the ED having three or more columns.
7. The method of claim 6 , wherein converting the plurality of potential column borders further comprises:
combining a portion of the column border group based on a horizontal overlap criterion, wherein the column border group identifies a region of the ED having three or more columns; and
merging the column border group with an adjacent column border group, wherein no intervening paragraph exists between the column border group and the adjacent column border group.
8. A non-transitory computer readable medium (CRM) storing computer readable program code for processing an electronic document (ED) to infer columns in the ED embodied therein, wherein the ED comprises a plurality of characters and the computer readable program code, when executed by a computer, comprises functionality for:
generating a mark-up version of the ED comprising text-layout attributes of the characters in the ED,
wherein
the characters are grouped into a plurality of paragraphs based on the text-layout attributes, and
each of the plurality of paragraphs corresponds to a paragraph bounding box surrounding a corresponding paragraph;
generating a plurality of border pieces by initiating a pair of left scan and right scan from each of the plurality of paragraph bounding boxes to identify any adjacent paragraph bounding box, at least one scan of the left scan and the right scan being initiated from said first one of the two adjacent paragraph bounding boxes to reach a second one of the two adjacent paragraph bounding boxes, wherein at least one border piece of the plurality of border pieces comprises:
a white space separating two adjacent paragraph bounding boxes of the plurality of paragraph bounding boxes in the horizontal direction;
a horizontal dimension equaling a separation between the two adjacent paragraph bounding boxes;
a vertical dimension equaling a height of a first one of the two adjacent paragraph bounding boxes;
a direction of said at least one scan of the left scan and the right scan; and
generating, based at least on the plurality of border pieces, a plurality of column borders for use in inferring the columns in the ED,
wherein at least one column comprises a vertically aligned portion of the plurality of paragraphs.
9. The CRM of claim 8 , wherein generating the plurality of column borders comprises:
generating a sorted list of the plurality of border pieces based on respective locations of the plurality of border pieces;
generating a plurality of potential column borders by initiating a pair of forward traversal and reverse traversal of the sorted list from each of the plurality of border pieces, wherein the pair of forward traversal and reverse traversal identifies any overlapping border pieces for combining into one of the plurality of potential column borders; and
converting the plurality of potential column borders into the plurality of column borders.
10. The CRM of claim 9 ,
wherein each of the plurality of border pieces corresponds to one of the plurality of paragraphs,
wherein each of the plurality of potential column borders is associated with a list of traversed border pieces based on a corresponding pair of forward traversal and reverse traversal, and
wherein converting the plurality of potential column borders comprises:
removing duplication among the plurality of potential column borders based at least on the list of traversed border pieces of each of the plurality of potential column borders.
11. The CRM of claim 9 , wherein converting the plurality of potential column borders comprises:
generating, from the plurality of potential column borders, a column border group comprising two or more potential column borders based on a vertical overlap criterion,
wherein the column border group identifies a region of the ED having three or more columns.
12. The CRM of claim 11 , wherein converting the plurality of potential column borders further comprises:
combining a portion of the column border group based on a horizontal overlap criterion, wherein the column border group identifies a region of the ED having three or more columns; and
merging the column border group with an adjacent column border group, wherein no intervening paragraph exists between the column border group and the adjacent column border group.
13. A system for processing an electronic document (ED) to infer columns in the ED, wherein the ED comprises a plurality of characters, the system comprising:
a memory; and
a computer processor connected to the memory that:
generates a mark-up version of the ED comprising text-layout attributes of the characters in the ED, wherein
the characters are grouped into a plurality of paragraphs based on the text-layout attributes, and
each of the plurality of paragraphs corresponds to a paragraph bounding box surrounding a corresponding paragraph;
generates a plurality of border pieces by initiating a pair of left scan and right scan from each of the plurality of paragraph bounding boxes to identify any adjacent paragraph bounding box, at least one scan of the left scan and the right scan being initiated from said first one of the two adjacent paragraph bounding boxes to reach a second one of the two adjacent paragraph bounding boxes, wherein at least one border piece of the plurality of border pieces comprises:
a white space separating two adjacent paragraph bounding boxes of the plurality of paragraph bounding boxes in the horizontal direction;
a horizontal dimension equaling a separation between the two adjacent paragraph bounding boxes;
a vertical dimension equaling a height of a first one of the two adjacent paragraph bounding boxes;
a direction of said at least one scan of the left scan and the right scan; and
generates, based at least on the plurality of border pieces, a plurality of column borders for use in inferring the columns in the ED,
wherein at least one column comprises a vertically aligned portion of the plurality of paragraphs.
14. The system of claim 13 , wherein generating the plurality of column borders comprises:
generating a sorted list of the plurality of border pieces based on respective locations of the plurality of border pieces;
generating a plurality of potential column borders by initiating a pair of forward traversal and reverse traversal of the sorted list from each of the plurality of border pieces, wherein the pair of forward traversal and reverse traversal identifies any overlapping border pieces for combining into one of the plurality of potential column borders; and
converting the plurality of potential column borders into the plurality of column borders.
15. The system of claim 14 ,
wherein each of the plurality of border pieces corresponds to one of the plurality of paragraphs,
wherein each of the plurality of potential column borders is associated with a list of traversed border pieces based on a corresponding pair of forward traversal and reverse traversal, and
wherein converting the plurality of potential column borders comprises:
removing duplication among the plurality of potential column borders based at least on the list of traversed border pieces of each of the plurality of potential column borders.
16. The system of claim 14 , wherein converting the plurality of potential column borders comprises:
generating, from the plurality of potential column borders, a column border group comprising two or more potential column borders based on a vertical overlap criterion,
wherein the column border group identifies a region of the ED having three or more columns.
17. The system of claim 16 , wherein converting the plurality of potential column borders further comprises:
combining a portion of the column border group based on a horizontal overlap criterion, wherein the column border group identifies a region of the ED having three or more columns; and
merging the column border group with an adjacent column border group, wherein no intervening paragraph exists between the column border group and the adjacent column border group.