Robust method for automatic reading of skewed, rotated or partially obscured characters
View Patent ↗A character reading technique recognizes character strings in grayscale images where characters within such strings have poor contrast, are variable in position or rotation with respect to other characters in the string, or where portions of characters in the string are partially obscured. The method improves classification accuracy by improving the robustness of the underlying correlation operation or the character design. Characters are divided into regions before performing correlations. Based upon the relative individual region results, region results are combined into a whole character result. Using the characters that are read, a running checksum is computed and, based upon the checksum result, characters are replaced to produce a valid result.
1. A method for designing a text symbol set to withstand partial obscuration wherein
a. an initial character set is created;
b. at least three regions of each character are selected;
c. for each character individually, one of the at least three regions of the character is obscured;
d. compute the correlation coefficients for all selected regions;
e. adjust the character set design to produce maximum overall character detection accuracy with one region obscured.
2. The method of claim 1 wherein the sub-regions are selected as pixels within the input image.
3. The method of claim 1 wherein the a-priori estimates of application regional obscuration probability characteristics are used in evaluating maximum overall character detection.
4. A method for region design in a regionalized character recognition system comprising:
a. selecting a character set design;
b. defining at least three sub-regions for each character template;
c. for each character, obscure one of the three regions;
d. compute the correlation coefficients for the entire character set
e. adjust the regions shape and overlap to produce maximum overall character detection.
5. The method of claim 4 wherein the weights are character feature template weights.
6. A method for weight learning in a regionalized character recognition system comprising:
a. selecting a character set design;
b. defining at least three sub-regions for each character template;
c. for each character, obscure one of the three regions;
d. compute the correlation coefficients for the entire character set
e. adjust the weights to produce maximum overall character detection.