Structural decomposition in handwriting

A computing device for processing lists in handwriting, comprising:

• a recognition module configured to recognise text by performing text handwriting recognition on a plurality of strokes of digital ink in an electronic document, said text comprising a plurality of successive text lines which can each be either a list item which is part of a list or a text item which is not part of a list;
• a classifier module configured to initially classify each recognised text line as a distinct text item and to perform a classification process comprising the following steps:
• a) pattern detection in each text line for determining whether said text line starts by a predetermined list symbol indicating that said text line is part of a list, and classifying each text line starting with a predetermined list symbol as a list item;
• b) in response to the step a), determining an item indentation of each text item with respect to a reference position and determining for each list item a text indentation representing the indentation of text comprised in said list item with respect to the reference position;
• c) in response to the step b), a merging process comprising:
• merging, as part of a same text item, each pair of consecutive text lines, in said text, formed by a first text line followed by a second text line if the first and second text lines of said pair are text items with a same item indentation and if there is not enough free space in the end of the first text line to accommodate a first word starting the second text line;
• and
• merging, as part of a same list item, each pair of consecutive text lines formed by a first text line followed by a second text line if the first text line is a list item and the second text line is a text item with an item indentation matching the text indentation of the first text line and if there is not enough free space in the end of the first text line to accommodate a first word starting the second text line;
• a model management module configured to generate, based on a result of the merging process, a text structure data model defining each text line as part of either a text item or a list item.