Claims Analysis
Total Claims: 9
Claim 1
Independent Claim
344 words
A method in a computer system of orienting an image for optical character recognition (OCR) comprising the following steps: a. subsampling a first portion of non-consecutive lines of a first image, the first image having a first orientation; b. storing the number of transitions for the first portion of non-consecutive lines of the first image in a first array; c. determining the number of transitions per line in the first array; d. scanning the first array and determining if there are a threshold number of sets of consecutive lines in the array having transitions, each of the sets of lines with transitions having a first number of lines with transitions, the first number in a first range between a minimum number of lines and a maximum number of lines; e. if there are a threshold number of sets of consecutive lines in the array with transitions, then indicating that the image is properly oriented, otherwise rotating the first image perpendicularly; f. if trial OCR has been performed upon the first image a second number of times then performing the remainder of OCR on the first image, placing the result in a first text file and terminating, otherwise continuing; g. performing trial OCR on a plurality of characters of the first image, and placing the result in the first text file; h. determining whether a first fraction of characters comprising the first text file is misclassified, and if so, then proceeding to step j; i. if a second fraction of characters comprising the first text file is misclassified, the second fraction being less than the first fraction, or the second fraction of the second text file is questionable, then proceeding to step j, otherwise performing the remainder of OCR on the first image and terminating; and j. determining whether the first image has been determined to be properly oriented or has been inverted, then rotating the first image perpendicularly, indicating that the first image is not right side up and proceeding to step f above, otherwise inverting the first image and proceeding to step f.
Claim 2
Independent Claim
344 words
A method in a computer system of orienting an image for optical character recognition (OCR) comprising the following steps: a. subsampling a first portion of non-consecutive lines of a first image, the first image having a first orientation; b. storing the number of transitions for the first portion of non-consecutive lines of the first image in a first array; c. determining the number of transitions per line in the first array; d. scanning the first array and determining if there are a threshold number of sets of consecutive lines in the array having transitions, each of the sets of lines with transitions having a first number of lines with transitions, the first number in a first range between a minimum number of lines and a maximum number of lines; e. if there are a threshold number of sets of consecutive lines in the array with transitions, then indicating that the image is properly oriented, otherwise rotating the first image perpendicularly; f. if trial OCR has been performed upon the first image a second number of times then performing the remainder of OCR on the first image, placing the result in a first text file and terminating, otherwise continuing; g. performing trial OCR on a plurality of characters of the first image, and placing the result in the first text file; h. determining whether a first fraction of characters comprising the first text file is misclassified, and if so, then proceeding to step j; i. if a second fraction of characters comprising the first text file is misclassified, the second fraction being less than the first fraction, or the second fraction of the second text file is questionable, then proceeding to step j, otherwise performing the remainder of OCR on the first image and terminating; and j. determining whether the first image has been determined to be properly oriented or has been inverted, then rotating the first image perpendicularly, indicating that the first image is not right side up and proceeding to step f above, otherwise inverting the first image and proceeding to step f.
Claim 3
Dependent Claim
21 words
The method of claim 1 wherein the steps of rotating the image perpendicularly comprise rotating the first image 90 degrees counterclockwise.
Claim 4
Dependent Claim
13 words
The method of claim 1 wherein the second number of times is four.
Claim 5
Dependent Claim
11 words
The method of claim 1 wherein the first fraction comprises 30%.
Claim 6
Dependent Claim
11 words
The method of claim 1 wherein the second fraction comprises 10%.
Claim 7
Dependent Claim
26 words
The method of claim 1 wherein the step of subsampling the image comprises regular sampling of non-consecutive positions along non-consecutive horizontal lines in the first image.
Claim 8
Dependent Claim
27 words
The method of claim 6 wherein the subsampling comprises selecting one of every eight pixels for each of one of every eight lines in the first image.
Claim 9
Dependent Claim
19 words
The method of claim 1 wherein the step of determining transitions comprises determining transitions between white and black pixels.