Deblurring and supervised adaptive thresholding for print-and-scan document image evaluation
A method is described to obtain a binary image from the print-and-scan process to best match the known original. A point-spread function (PSF) of the PAS process is first obtained from its knife-edge responses, and deblurring is carried out on the scanned images using deconvolution. After image deskewing and preliminary registration, a supervised adaptive thresholding procedure is utilized to binarize the scanned image such that a measure of difference (e.g. the Euclidean distance) between the original and binarized images is minimized. The supervised adaptive thresholding procedure divides the scanned images into many rectangular sub-images. Otsu's method is used to find a starting threshold for each scanned sub-image. An optimal threshold is found around the Otsu's threshold via iterative search to minimize the measure of difference between the original sub-image and scanned sub-image. The sub-images are binarized using the optimal threshold. This method may be used in document authentication.
1. A method implemented in a data processing system for processing a hardcopy document, the hardcopy document having been printed by a printer based on a original image, the original image being a binary digital image, comprising:
(a) generating a grayscale scanned image from the hardcopy document using an imaging device;
(b) obtaining the original image;
(c) dividing the original image into a plurality of original sub-images;
(d) for each original sub-image,
(d1) using template matching to find a scanned sub-image in the scanned image corresponding to the original sub-image;
(d2) obtaining an initial threshold for binarizing the scanned sub-image; and
(d3) obtaining an optimal threshold for binarizing the scanned sub-image using the initial threshold and an iterative search, and binarizing the scanned sub-image using the optimal threshold to generate an optimum binarized scanned sub-image that minimizes a measure of difference between the original sub-image and the binarized scanned sub-image; and
(e) generating a binarized scanned image by combining the optimum binarized scanned sub-images generated in step (d).
2. The method of claim 1 , further comprising, before step (d):
(f) deblurring the scanned images by applying a deconvolution method to the scanned images using a point-spread function characteristic of the printer and the imaging device.
3. The method of claim 2 , wherein the point-spread function is a two-dimensional point-spread function, wherein the method further comprises, before step (f),
(g) calculating the two-dimensional point-spread function, comprising:
(g1) generating a test binary image containing sharp edges in two non-parallel directions;
(g2) printing the test image using the printer;
(g3) generating a grayscale scanned test image from the printed test image using the imaging device;
(g4) calculating edge spread functions in two directions using pixel intensities of the scanned test image in vicinities of the sharp edges; and
(g5) calculating the two-dimensional point-spread function from the edge spread functions.
4. The method of claim 1 , further comprising, after step (e):
(f) applying morphological operations with asymmetric kernels to the binarized scanned image.
5. The method of claim 1 , further comprising, before step (d):
(h) performing deskewing and registration of the scanned image to generate a deskewed and registered scanned image,
wherein step (d) is performed on the deskewed and registered scanned image.
6. The method of claim 5 , wherein step (h) comprises:
(h1) performing preliminary thresholding on the scanned image to generate a preliminary binary image;
(h2) extracting a representative pixel from each connected image component in the preliminary binary image to generate a simplified image;
(h3) calculating a skew angle of the simplified image;
(h4) calculating a deskewed scanned image from the scanned image using the skew angle; and
(h5) performing template matching between the original image and the deskewed scanned image to generate the deskewed and registered scanned image.
7. The method of claim 1 , wherein step (b) comprises obtaining the original image from a database.
8. A computer program product comprising a computer usable non-transitory medium having a computer readable program code embedded therein for controlling a data processing apparatus, the computer readable program code configured to cause the data processing apparatus to execute a process for processing a hardcopy document, the hardcopy document having been printed by a printer based on a digital original image, the original image being a binary image, the process comprising:
(a) generating a grayscale scanned image from the hardcopy document using an imaging device;
(b) obtaining the original image;
(c) dividing the original image into a plurality of original sub-images;
(d) for each original sub-image,
(d1) using template matching to find a scanned sub-image in the scanned image corresponding to the original sub-image;
(d2) obtaining an initial threshold for binarizing the scanned sub-image; and
(d3) obtaining an optimal threshold for binarizing the scanned sub-image using the initial threshold and an iterative search, and binarizing the scanned sub-image using the optimal threshold to generate an optimum binarized scanned sub-image that minimizes a measure of difference between the original sub-image and the binarized scanned sub-image; and
(e) generating a binarized scanned image by combining the optimum binarized scanned sub-images generated in step (d).
9. The computer program product of claim 8 , wherein the process further comprises, before step (d):
(f) deblurring the scanned images by applying a deconvolution method to the scanned images using a point-spread function characteristic of the printer and the imaging device.
10. The computer program product of claim 9 , wherein the point-spread function is a two-dimensional point-spread function, wherein the process further comprises, before step (f),
(g) calculating the two-dimensional point-spread function, comprising:
(g1) generating a test binary image containing sharp edges in two non-parallel directions;
(g2) printing the test image using the printer;
(g3) generating a grayscale scanned test image from the printed test image using the imaging device;
(g4) calculating edge spread functions in two directions using pixel intensities of the scanned test image in vicinities of the sharp edges; and
(g5) calculating the two-dimensional point-spread function from the edge spread functions.
11. The computer program product of claim 8 , wherein the process further comprises, after step (e):
(f) applying morphological operations with asymmetric kernels to the binarized scanned image.
12. The computer program product of claim 8 , wherein the process further comprises, before step (d):
(h) performing deskewing and registration of the scanned image to generate a deskewed and registered scanned image,
wherein step (d) is performed on the deskewed and registered scanned image.
13. The computer program product of claim 12 , wherein step (h) comprises:
(h1) performing preliminary thresholding on the scanned image to generate a preliminary binary image;
(h2) extracting a representative pixel from each connected image component in the preliminary binary image to generate a simplified image;
(h3) calculating a skew angle of the simplified image;
(h4) calculating a deskewed scanned image from the scanned image using the skew angle; and
(h5) performing template matching between the original image and the deskewed scanned image to generate the deskewed and registered scanned image.
14. The computer program product of claim 8 , wherein step (b) comprises obtaining the original image from a database.