Image rectification using text line tracks

A method for rectifying an input digital image including warped textual lines, the method implemented at least in part by a data processing system and comprising:

• receiving an input digital image;
• analyzing the input digital image to determine a plurality of text line traces that track along with the warped textual lines in a substantially parallel fashion;
• subdividing the input digital image into a plurality of vertical slices, each vertical slice having associated first and second slice edges and an associated slice width;
• forming vertical slice deformation models for a plurality of the vertical slices by:
• identifying a plurality of text line traces that cross a particular vertical slice, each of the identified text line traces having a first-edge vertical coordinate corresponding to the location where the text line trace crosses the first slice edge and a second-edge vertical coordinate corresponding to the location where the text line trace crosses the second slice edge of the vertical slice;
• and
• forming the vertical slice deformation model for the particular vertical slice by fitting a parametric function that predicts the second-edge vertical coordinates as a function of the first-edge vertical coordinates;
• forming a plurality of streamlines responsive to the vertical slice deformation models, wherein the streamlines represent a deformation of the warped textual lines in the input digital image;
• forming a global deformation function by interpolating between the streamlines;
• forming a rectified image by rectifying the input digital image responsive to the global deformation function;
• and
• storing the rectified image in a processor-accessible memory.