Multi-font optical character recognition apparatus

• US 4,097,847 A
• Filed: 05/14/1975
• Issued: 06/27/1978
• Est. Priority Date: 07/10/1972
• Status: Expired due to Term

• Current Assignee: Scan-Optics LLC
• Original Assignee: Scan-Optics, Incorporated (Scan-Optics LLC)
• Inventors: Jackson, Stephen F., Forsen, George E.
• Primary Examiner(s): Boudreau, Leo H.
• Application Number:
  US05/577,151
• Time in Patent Office: 1,140 Days
• Field of Search: 340/146.3 AE, 340/146.3 H, 340/146.3 AC, 340/146.3 R, 340/146.3 AG
• US Class Current: 382/199
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06V 10/10 Image acquisition document image scanning and transmission H04N1/00; control of digital cameras H04N23/60
• CPC Class Codes: G06V 10/12 Details of acquisition arrangements; Constructional details thereof
• CPC Class Codes: G06V 10/20 Image preprocessing
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/146 Aligning or centring of the image pick-up or image-field
• CPC Class Codes: G06V 30/162 Quantising the image signal
• CPC Class Codes: G06V 30/18 Extraction of features or characteristics of the image
• CPC Class Codes: G06V 30/182 by coding the contour of the pattern
• CPC Class Codes: G06V 30/192 using simultaneous comparisons or correlations of the image signals with a plurality of references

Abstract

As described herein, a program controlled image dissector tube scans the printed information recorded on a storage medium in a first raster to provide analog information signals representative of segments of the recorded information and in a second raster to provide analog information signals representative of all the recorded information. If, as determined by the analog information signals representative of segments of the recorded information, the printed information is legible, the analog information signals representative of all the printed information are converted into digital form and scaled, if necessary, to reduce the digital character data to a predetermined scan length. Selected arrays of the digital character data are then scanned to develop black or white signals about a center bit within such arrays. The black or white signals are, in turn, accumulated to provide an array of black and white signals representative of an entire character. The contour of the array of black and white signals is then traced to develop contour digital data representative of incremental moves about the perimeter of the character. The extent and direction of such incremental moves are then accumulated to develop parameters. The parameters are then selectively combined and the areas bounded by such parameters calculated to develop feature indication signals. The feature indication signals are then accumulated, compared against preselected recognition thresholds and classified as a particular character.

Claims

1. Multi-font optical character recognition apparatus for converting character representative symbols located on a storage medium in the form of a two-dimensional field of brightness values into a digital code representative of said symbol, said apparatus including scanning means for optically scanning the storage medium to generate analog signals representative of the brightness values, contrast means for measuring the difference between the analog signals produced by the various brightness values, means for converting the analog signals into one bit digital data signals denoting either blackness for the symbol or whiteness for the background in dependence on the measurement of the contrast means, perimeter trace means for developing contour digital data representative of incremental movements between data points about the perimeter of the symbol, and recognition means for converting the contour data into a digital code representing the scanned symbol, wherein the improvement comprises a perimeter tracing means including:

a character memory for storing the one bit digital data signals in predetermined arrays representing corresponding segmental areas of the scanned symbol,

address means for addressing particular data bits in the character memory,

a perimeter trace table memory for storing a search sequence for finding the perimeter of the character, and

control means for causing the address means to follow the search sequence stored in the perimeter trace table so that the perimeter of the symbol stored in the character memory is searched for in a 5 ×

5 array of a portion of the character memory according to the following sequence;

row 2, col. 1;

row 2, col. 2;

row 2, col. 3;

row 1, col. 2;

row 1, col. 3;

row 1, col. 4;

row 2, col. 4;

row 3, col. 4;

row 2, col. 5;

row 3, col. 5;

row 4, col. 5;

row 4, col. 4;

row 4, col. 3;

row 5, col. 4;

row 5, col. 3;

row 5, col. 2;

row 4, col. 2;

row 3, col. 2;

row 4, col. 1;

row 3, col. 1;

row 2, col. 1;

stopping in the search whenever a part of the perimeter is found, and beginning a new search with a shifted 5 ×

5 array.

2. Multi-font optical character recognition apparatus according to claim 1 wherein the means for converting the analog signals includes means for digitizing the analog signals into multi-bit digital data signals representative of the amplitude of the analog signals and a preprocessing means for conditioning the multi-bit digital data signals, including means for accumulating the digital information into predetermined arrays and means for changing the information into the one bit digital data denoting either blackness or whiteness.

3. A multi-font optical character recognition apparatus according to claim 2 further including a scaling means in the preprocessing means for scaling the multi-bit digital data signals to a predetermined number of new digital data signals representative of the brightness in selected segments of the scanned storage medium comprising register means for storing the digital data signals developed from a first complete line in the scan of the entire storage medium and means for comparing the stored digital data signals with the digital data signals developed from a second complete line in the raster scan of the entire storage medium for blackness values and transmitting the multi-bit digital data having the greatest blackness values to the accumulating means of said preprocessing means.

4. A multi-font optical character recognition apparatus according to claim 1 wherein said contrast measurement means includes a control means for said image scanning means, said control means supplying first vertical and horizontal scanning signals to the image scanning means to generate a preliminary raster scan of selected portions of said storage medium and develop information representative of the legibility of the character representative symbols based on the difference in the analog signals produced from the symbol and its background, and second vertical and horizontal scanning signals to the image scanning means to generate a raster scan of the entire storage medium to develop information representative of the recorded character representative symbols.

5. A multi-font optical character recognition apparatus according to claim 3 wherein the accumulating means of the preprocessing means comprises register means for sequentially accumulating the scaled digital data signals into sequential arrays corresponding to the scanning of a selected area on the storage medium by the scanning means and wherein the means for changing comprises means for sampling the arrays in, through and about the centers of the arrays and comparing the center bit values with the values of the bits surrounding the center bit in said arrays for blackness to classify the center bit values as either black or white.

6. Contrast means for developing information representative of the legibility of symbols in a multi-font optical character recognition system, in which character symbols in the form of fields of brightness values recorded on a storage medium are scanned to develop analog information signals representative of the symbols, comprising:

means for vertical and horizontal pre-scanning of the symbols in selected areas according to a predetermined pattern prior to scanning to develop the analog information signals representative of the symbols;

means for measuring the differences between analog signals generated during the pre-scanning; and

means for determining if the symbols can be read and the best speed at which to scan it on the basis of the differences between analog signals generated during the pre-scanning.

7. Image enhancement means for converting multi-bit digital signals into one-bit representations denoting either blackness or whiteness in a multi-font optical character recognition system in which character symbols in the form of a field of brightness values recorded on a storage medium are converted into a predetermined array of digital signals for each segmental area of the symbol, comprising:

means for storing the digital signals that represent a 3 ×

3 array centered about the multi-bit digital signal to be enhanced;

means for generating a first signal if the multi-bit digital signal to be enhanced is less than a first reference level;

means for generating a second signal if the multi-bit digital signal to be enhanced is less than a second reference level;

means for generating eight sum signals equal to the summation of the multi-bit digital values in the horizontal, vertical and diagonal rows of the 3 ×

3 array;

means for generating a third signal if one of the four sum signals representative of the sum of the multi-bit digital data in the horizontal, vertical and diagonal lines passing through the 3 ×

3 array center, is less than a third reference level;

p1 means for generating a fourth signal if any one of the four sum signals representative of the sum of the multi-bit digital data in the horizontal, vertical and diagonal lines passing through the 3 ×

3 array center are less positive than any one of the four sum signals representative of the sum of the multi-bit digital data in the horizontal lines about and below the 3 ×

3 array center and the vertical lines to the left and right of the 3 ×

3 array center; and

means for setting the one-bit representation of the center segmental area to be enhanced to a blackness signal whenever the first signal is present or whenever the second, third and fourth signals are present simultaneously.

8. The system of claim 7 further including perimeter trace processing means for developing contour signals from the one-bit representations of the symbol by searching for the perimeter of the symbol in a 5 ×

5 matrix of a portion of the array of one-bit representations according to the following digital search sequence;

row 2, col. 1;

row 2, col. 2;

row 2, col. 3;

row 1, col. 2;

row 1, col. 3;

row 1, col. 4;

row 2, col. 4;

row 3, col. 4;

row 2, col. 5;

row 3, col. 5;

row 4, col. 5;

row 4, col. 4;

row 4, col. 3;

row 5, col. 4;

row 5, col. 3;

row 5, col. 2;

row 4, col. 2;

row 3, col. 2;

row 4, col. 1;

row 3, col. 1;

row 2, col. 1; and

stopping in the search and returning to row 2, col. 1, whenever a one-bit representation of blackness is found.