Image resizing for optical character recognition in portable reading machine
A reading machine that operates in various modes includes image correction processing is described. The reading device pre-processes an image for optical character recognition by receiving the image and determining whether text in the image is too large or small for optical character recognition processing by determining that text height falls outside of a range in which optical character recognition software will recognize text in a digitized image. If necessary the image is resized according to whether the text is too large or too small.
1. A computer program product tangibly embodied on a computer readable hardware storage device, the computer program product for processing an image to prepare the image for optical character recognition, the computer program product comprising instructions for causing a machine to:
receive a digitized image for optical character recognition processing;
apply optical character recognition processing to the digitized image;
evaluate whether text in the digitized image is too large or too small for optical character recognition processing by instructions to:
determine that text height falls outside of a range in which optical character recognition software will recognize text because the text is too large; and
determine that text height falls outside of a range in which optical character recognition software will recognize text because the text is too small;
resize the received digitized image and concomitant therewith text in the received digitized image according to whether the evaluation of the text to bring the text height into the range; and
cause reapplication of optical character recognition to be applied to the received, resized digitized image.
2. The computer program product of claim 1 wherein the range is approximately 20 to 170 pixels in height.
3. The computer program product of claim 1 further comprising:
applying the Hough transform to the digitized image to provide an estimate of text size in the image.
4. The computer program product of claim 1 wherein if the text is determined to be too small, instructions to resize, comprise instructions to:
mathematically double the size of the digitized image and corresponding text in the digitized image and add in missing pixels using an interpolation process.
5. The computer program product of claim 1 wherein if the text is determined to be too large, instructions to resize further comprise instructions to:
apply decimation to the digitized image to reduce the size of the text in the digitized image.
6. The computer program product of claim 5 wherein decimation ratios are determined by the largest expected size of the text.
7. The computer program product of claim 5 wherein decimation ratios are determined such that characters in the digitized image are at a pixel height for optical character recognition of the digitized image.
8. The computer program product of claim 5 wherein the received image is a first received image and the program processes plural images including the first received image to provide plural, received re-sized digitized images at different decimation ratios and with the different decimation ratios chosen to provide overlap in the text in images with the different decimation ratios.
9. The computer program product of claim 1 wherein a plurality of different re-sizings of the received, resized digitized image are provided from the instructions to resize, and the instructions apply optical character recognition (OCR) to each re-sizing; and
determine a quality measure of the optical character recognition (OCR).
10. The computer program product of claim 9 wherein the instructions apply OCR to each resizing until the determined quality measure of optical character recognition is sufficient.
11. A computer implemented method for processing an image for optical character recognition, the method comprising:
receiving in a computing device a digitized image for optical character recognition processing;
applying optical character recognition processing to the digitized image;
evaluating by the computing device whether text in the digitized image is too large for optical character recognition processing by determining that text height falls outside of an upper portion of a range in which optical character recognition software will recognize text;
resizing by the computing device the received digitized image and concomitant therewith text in the received digitized image when the computing device determines that the text is too large for suitable optical character recognition to bring the text height with the upper portion of the range;
causing reapplication of optical character recognition to be applied to the received, resized digitized image.
12. The computer implemented method of claim 11 wherein the range is approximately 20 to 170 pixels in height, the method further comprising:
evaluating by the computing device whether text in the digitized image is too small for optical character recognition processing by the computing device determining that text height falls outside of a lower portion of the range in which optical character recognition software will recognize text; and
resizinq by the computing device the received digitized image and concomitant therewith text in the received digitized image according to whether the evaluating of text determines that the text is too large or too small for suitable optical character recognition to bring the text height into the range.
13. The computer implemented method of claim 11 further comprising:
applying the Hough transform to the digitized image to provide an estimate of text size in the image.
14. The computer implemented method of claim 12 wherein if the text is determined to be too small, instructions to resize, comprise instructions to:
mathematically double the size of the digitized image by the computing device and corresponding text in the digitized image and add in missing pixels using an interpolation process.
15. The computer implemented method of claim 11 wherein if the text is determined to be too large, instructions to resize further comprise instructions to:
apply decimation to the digitized image to reduce the size of the text in the digitized image.
16. The computer implemented method of claim 15 wherein decimation ratios are determined by the largest expected size of the text.
17. The computer implemented method of claim 15 wherein decimation ratios are determined such that characters in the digitized image are at a pixel height for optical character recognition of the digitized image.
18. The computer implemented method of claim 15 wherein the received image is a first received image and the program processes plural images including the first received image to provide plural re-sized images at different decimation ratios and with the different decimation ratios chosen to provide overlap in the text in images with the different decimation ratios.
19. The computer implemented method of claim 11 wherein the received image is a first received image, the method further comprises processing plural images including the first received image to provide plural re-sized images at different decimation ratios with the different decimation ratios chosen to provide overlap in the text in the digitized images, and
determining a quality measure of the optical character recognition;
applying optical character recognition (OCR) to each re-sizing until the quality measure of OCR is sufficient.
20. The computer implemented method of claim 19 , further comprising instructions to:
applying optical character recognition (OCR) to the received, resized digitized image.
21. Apparatus comprising:
a computer device for processing an image for optical character recognition, the device comprising:
a processor; and
circuitry to:
receive a digitized image for processing;
apply optical character recognition processing to the digitized image;
evaluate whether text in the digitized image is too large for optical character recognition processing by the circuitry configured to:
determine that text height falls outside of an upper portion of a range in which optical character recognition software will recognize text;
evaluate whether text in the digitized image is too small for optical character recognition processing by the circuitry configured to:
determine that text height falls outside of a lower portion of the range in which optical character recognition software will recognize text; and
resize the received image and concomitant therewith text in the received digitized image according to the evaluation made by the circuitry of the text to bring the text height into the range; and
cause reapplication of optical character recognition to be applied to the received, resized digitized image.
22. The apparatus of claim 21 wherein the range is approximately 20 to 170 pixels in height.
23. The apparatus of claim 21 further comprising:
applying the Hough transform to the digitized image to provide an estimate of text size in the image.
24. The apparatus of claim 21 wherein if the text is determined to be too small, instructions to resize, comprise instructions to:
mathematically double the size of the digitized image and corresponding text in the digitized image and add in missing pixels using an interpolation process.
25. The apparatus of claim 21 wherein if the text is determined to be too large, instructions to resize further comprise instructions to:
apply decimation to the digitized image to reduce the size of the text in the digitized image.
26. The apparatus of claim 25 of wherein decimation ratios are determined by the largest expected size of the text.
27. The apparatus of claim 25 wherein decimation ratios are determined such that characters in the digitized image are at a pixel height for optical character recognition of the digitized image.
28. The apparatus of claim 25 wherein the received image is a first received image and the program processes plural images including the first received image to provide plural, received re-sized digitized images at different decimation ratios and with the different decimation ratios chosen to provide overlap in the text in images with the different decimation ratios.
29. The apparatus of claim 21 wherein a plurality of different re-sizings of the received, resized digitized image are provided from the instructions to resize, and the instructions apply optical character recognition (OCR) to each re-sizing; and
determine a quality measure of the optical character recognition.
30. The apparatus of claim 29 wherein the instructions apply OCR to each resizing until the determined the quality measure of the optical character recognition is sufficient.