Method for assessing the quality of an image of a document

1. A computer-implemented quality assessment method for assessing the quality of a document image to be processed in an OCR system prior to OCR processing, the method comprising the steps of, performed in a computerized system:

a) processing the document image to a text image comprising a number of text blobs, each text blob comprising a region of connected pixels having the same value;

b) classifying the text blobs in the text image in a first and second type of text blobs based on a calculation as to whether they will belong to a foreground layer in OCR processing or to a background layer in OCR processing; and

c) generating a quality value for the document image prior to OCR processing based on the classified text blobs;

wherein step b) further comprises:

b1) calculating a text compression cost and a picture compression cost for each text blob;

b2) calculating a ratio of the text compression cost and the picture compression cost for each text blob;

b3) comparing said ratio to a predetermined threshold to determine if said ratio is less than the predetermined threshold;

b4) classifying said text blob as a second type blob if the ratio is less than the predetermined threshold; and

b5) classifying said text blob as a first type blob if the ratio is not less than the predetermined threshold.

2. The computer-implemented quality assessment method according to claim 1 , characterized in that step b1) comprises:

b11a) calculating a background compression cost for each text blob;

b12a) calculating a foreground compression cost for each text blob; and

b13a) calculating a mask compression cost for each text blob.

3. The computer-implemented quality assessment method according to claim 2 , characterized in that step b1) further comprises:

b14a) adding the foreground compression cost and the mask compression cost to calculate the text compression cost for each text blob; and

b15a) equating the picture compression cost with the background compression cost for each text blob.

4. The computer-implemented quality assessment method according to claim 2 , characterized in that

step b11a) comprises calculating a sum of squares of grayscale differences between a target pixel in said text blob and its eight touching pixels;

step b12a) comprises calculating a sum of squares of grayscale differences between a grayscale value of a target pixel in said text blob and an average grayscale value of pixels in the text blob; and

step b13a) comprises calculating a perimeter of said text blob.

5. The computer-implemented quality assessment method according to claim 1 , characterized in that step b1) comprises:

b11b) calculating a background compression cost for each text blob;

b12b) calculating a surrounding compression cost for each text blob;

b13b) calculating a foreground compression cost for each text blob; and

b14b) calculating a mask compression cost for each text blob.

6. The computer-implemented quality assessment method according to claim 5 , characterized in that step b1) further comprises:

b15b) adding the foreground compression cost, the mask compression cost and the surrounding compression cost to calculate the text compression cost for each text blob; and

b16b) equating the picture compression cost with the background compression cost for each text blob.

7. The computer-implemented quality assessment method according to claim 5 , characterized in that

step b11b) comprises calculating a sum of squares of grayscale differences between surrounding pixels and an average color of the surrounding pixels multiplied by a pre-set factor, the surrounding pixels being background pixels near an edge of the text blob;

step b12b) comprises calculating a sum of squares of grayscale differences between a grayscale value of a target pixel in said text blob and an average grayscale value of pixels in the text blob; and

step b13b) comprises calculating a perimeter of said text blob.

8. The computer-implemented quality assessment method according to claim 1 , characterized in that step c) comprises calculating a ratio of the number of first type blobs to the total number of text blobs.

9. The computer-implemented quality assessment method according to claim 1 , characterized in that the document image is a color image and the method further comprises, prior to step a), processing the color image to form a grayscale image.

10. The computer-implemented quality assessment method according to claim 1 , characterized in that step a) further comprises:

a1) binarizing the document image to form a binary image; and

a2) separating text elements from the binary image to form the text image.

11. The computer-implemented quality assessment method according to claim 10 , characterized in that step a2) comprises:

a21) identifying blobs in the binary image; and

a22) classifying each blob as one of a pictorial element and a textual element.

12. The computer-implemented quality assessment method according to claim 11 , characterized in that step a22) comprises:

classifying each blob as a pictorial element if the area of said blob is too large or too small compared to predefined thresholds.

13. The computer-implemented quality assessment method according to claim 11 , characterized in that step a22) further comprises:

calculating a stroke of each blob; and

classifying each blob as a pictorial element if the stroke of said blob is too large compared to a predefined threshold.

14. The computer-implemented quality assessment method according to claim 11 , characterized in that step a22) further comprises:

calculating a width and a height of each blob; and

classifying each blob as a pictorial element if at least one of the width and the height of said blob is too large compared to a predefined threshold.

15. The computer-implemented quality assessment method according to claim 1 , wherein the first type of blob is a text-compression-prone blob and the second type of blob is a picture-compression-prone blob.