Method of multi-font template enhancement by pixel weighting
First Claim
1. The method of matching a pixel version of an unknown input symbol with a library of enhanced pixel templates by recognition enhancement of a library of L unenhanced pixel images (I1 I2 I3 . . . Ij . . . IL) with respect to a pre-existing group of G classes of symbols (S1 S2 S3 . . . Si . . . SG) at least some of which include multiple pixel versions of the same symbol, for providing a library of G recognition enhanced pixel templates (T1 T2 T3 . . . Ti . . . TG) one enhanced template corresponding to each of the symbol classes, comprising the steps of:
- providing a library of L unenhanced pixel images (I1 I2 I3 . . . Ij . . . IL) to be enhanced to provide a library of G recognition enhanced pixel templates (T1 T2 T3 . . . Ti . . . TG);
providing a pre-existing group of G symbol classes (S1 S2 S3 . . . Si . . . SG) with the multiple pixel versions therein, a total of V pixel versions;
comparing each of the V pixel versions with each of the L pixel images to obtain V×
L comparisons forming V sets of L comparisons (C1 C2 C3 . . . Cj . . . CL), one set of L comparisons for each of the V pixel versions, each set of comparisons having a comparison Ci for each pixel image Ij ;
identifying a primary comparison C* from the L comparisons within each of the V sets of comparisons having the closest comparison with the pixel version for that set of comparisons forming a collection of V primary comparisons C* (C1 * C2 * C3 * . . . Ci * . . . Cv *);
identifying a secondary comparison C** from the L-1 remaining comparisons within each of the V sets of comparisons having the next closest comparison with the pixel version for that set of comparisons forming a collection of V secondary comparisons C** (C1 ** C2 ** C3 ** . . . Ci ** CV **), to provide V pairs of identified comparisons C* and C**, one pair from each of the V sets of comparisons;
determining V recognition margins (M1 M2 M3 . . . Mi . . . MV), one recognition margin between each pair of identified comparisons C* and C**;
selecting the single pair of identified comparisons C* and C** having the smallest recognition margin M* of all of the V pairs of identified comparisons from the V sets of comparisons;
identifying a single symbol within a class S* corresponding to the selected single comparison C*, and determining if the class S* has multiple pixel versions, and excluding the multiple pixel versions, if any, in the class S* from the remaining steps leaving the identified single symbol;
identifying the single pair of pixel images I* and I** corresponding to the selected single pair of identified comparisons C* and C**;
weighting certain pixels of the closest pixel image I* and the next closest pixel image I** corresponding to the selected single pair of identified comparisons C* and C** in order to incrementally increase the recognition margin M* therebetween causing the pixel image I* to become the closest enhanced pixel template T* and the pixel image I** to become the next closest enhanced pixel template T**;
iterating the comparing, identifying, determining, selecting, and weighting steps until the library of pixel images has become the library of G recognition enhanced symbol templates (T1 T2 T3 . . . Ti . . . TG) which have been recognition enhanced with respect to the pre-existing group of G pixel symbols (S1 S2 S3 . . . Si . . . SG), at least some templates of which have weighted pixel aberrations not present in the pixel version within each of the pre-existing group of G corresponding symbol classes; and
matching a pixel version of an unknown input symbol of the group of G symbol classes (S1 S2 S3 . . . Si . . . SG) with the library of enhanced templates (T1 T2 T3 . . . Ti . . . TG) by comparing the pixel version with each of the enhanced pixel templates in the library of enhanced templates and selecting the enhanced template with the closest comparison.
1 Assignment
0 Petitions
Accused Products
Abstract
A library of L unenhanced images (pixel bit maps) is enhanced for optical character recognition (OCR) with respect to a pre-existing group of G input symbol classes for creating a library of G recognition enhanced templates (pixel bit maps) of the G input symbol classes. The enhancement is accomplished by comparing each image of the library with each pixel version of each symbol class (forming V sets of L comparisons), and weighting the images with the highest potential for confusion. Divergent pixel version are further divided into sub-classes, each with a sub-image which evolves into a sub-template. A primary comparison C* and a secondary comparison C** are identified from the L comparisons within each of the V sets of comparisons. A recognition margin is determined between each pair of identified comparisons C* and C**. The single pair of identified comparisons C* and C** is selected forming the smallest recognition margin M*. The single pair of images Ij * and Ij ** underlying the pair of identified comparisons C* and C** is identified. Certain pixels of the closest pixel image Ij * and the next closest pixel image Ij ** are weighted in order to increase the recognition margin M* therebetween. The steps are iterated until the library of pixel images (and sub-images) has become a library of enhanced symbol templates (and sub-templates). The weighted pixel aberrations generated in the enhanced templates (and sub-templates) are not present in the corresponding symbol of the pre-existing group of G input symbol classes (and sub-classes).
9 Citations
31 Claims
-
1. The method of matching a pixel version of an unknown input symbol with a library of enhanced pixel templates by recognition enhancement of a library of L unenhanced pixel images (I1 I2 I3 . . . Ij . . . IL) with respect to a pre-existing group of G classes of symbols (S1 S2 S3 . . . Si . . . SG) at least some of which include multiple pixel versions of the same symbol, for providing a library of G recognition enhanced pixel templates (T1 T2 T3 . . . Ti . . . TG) one enhanced template corresponding to each of the symbol classes, comprising the steps of:
-
providing a library of L unenhanced pixel images (I1 I2 I3 . . . Ij . . . IL) to be enhanced to provide a library of G recognition enhanced pixel templates (T1 T2 T3 . . . Ti . . . TG); providing a pre-existing group of G symbol classes (S1 S2 S3 . . . Si . . . SG) with the multiple pixel versions therein, a total of V pixel versions; comparing each of the V pixel versions with each of the L pixel images to obtain V×
L comparisons forming V sets of L comparisons (C1 C2 C3 . . . Cj . . . CL), one set of L comparisons for each of the V pixel versions, each set of comparisons having a comparison Ci for each pixel image Ij ;identifying a primary comparison C* from the L comparisons within each of the V sets of comparisons having the closest comparison with the pixel version for that set of comparisons forming a collection of V primary comparisons C* (C1 * C2 * C3 * . . . Ci * . . . Cv *); identifying a secondary comparison C** from the L-1 remaining comparisons within each of the V sets of comparisons having the next closest comparison with the pixel version for that set of comparisons forming a collection of V secondary comparisons C** (C1 ** C2 ** C3 ** . . . Ci ** CV **), to provide V pairs of identified comparisons C* and C**, one pair from each of the V sets of comparisons; determining V recognition margins (M1 M2 M3 . . . Mi . . . MV), one recognition margin between each pair of identified comparisons C* and C**; selecting the single pair of identified comparisons C* and C** having the smallest recognition margin M* of all of the V pairs of identified comparisons from the V sets of comparisons; identifying a single symbol within a class S* corresponding to the selected single comparison C*, and determining if the class S* has multiple pixel versions, and excluding the multiple pixel versions, if any, in the class S* from the remaining steps leaving the identified single symbol; identifying the single pair of pixel images I* and I** corresponding to the selected single pair of identified comparisons C* and C**; weighting certain pixels of the closest pixel image I* and the next closest pixel image I** corresponding to the selected single pair of identified comparisons C* and C** in order to incrementally increase the recognition margin M* therebetween causing the pixel image I* to become the closest enhanced pixel template T* and the pixel image I** to become the next closest enhanced pixel template T**; iterating the comparing, identifying, determining, selecting, and weighting steps until the library of pixel images has become the library of G recognition enhanced symbol templates (T1 T2 T3 . . . Ti . . . TG) which have been recognition enhanced with respect to the pre-existing group of G pixel symbols (S1 S2 S3 . . . Si . . . SG), at least some templates of which have weighted pixel aberrations not present in the pixel version within each of the pre-existing group of G corresponding symbol classes; and matching a pixel version of an unknown input symbol of the group of G symbol classes (S1 S2 S3 . . . Si . . . SG) with the library of enhanced templates (T1 T2 T3 . . . Ti . . . TG) by comparing the pixel version with each of the enhanced pixel templates in the library of enhanced templates and selecting the enhanced template with the closest comparison. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31)
-
Specification