Character recognition techniques

Claim 3

The method of claim 1 wherein said step of quantizing comprises the steps of: developing a first field array of first binary levels when those voltages included in said array of voltages exceed a first quantizing level and second binary levels when those voltages included in said array of voltages are less than said first quantizing level; developing a second field array of first binary levels when those voltages included in said array of voltages exceed a second quantizing level and second binary levels when those voltages included in said array of voltages are less than said second quantizing level, said second quantizing level being less than said first quantizing level; and developing a third field array of first binary levels when those voltages included in said array of voltages exceed a third quantizing level and second binary levels when those voltages included in said array of voltages are less than said third quantizing level, said third quantizing level being less than said second quantizing level.

Claim 4

The method of claim 2 wherein said steps of developing a first field array, developing a second field array and developing a third field array each include the step of arranging said first and second binary levels in rows and columns whereby said first field array represents a normalized character formed of relatively thick strokes, said second field array represents a character formed of relatively normal strokes and said third field array represents a normalized character formed of relatively thin strokes.

Claim 5

The method of claim 3 wherein said step of sequentially comparing comprises the steps of: simultaneously detecting the coincidence of selected ones of said first and second binary levels of a given row of successive field arrays with corresponding binary levels that uniquely form each of said predetermined binary characters; and repeating the last-mentioned step in succession for each of the remaining rows of said field arrays.

Claim 6

The method of claim 4 wherein said step of sequentially comparing further comprises the step of indicating the coincidence of each row of each field array with one of said predetermined binary characters whereby the coinciding row is characterized as said one binary character.

Claim 9

Character recognizing apparatus comprising: analyzing means for generating binary signals representing the information content of a geometric pattern corresponding to a presented character; storage means coupled to said analyzing means for storing said binary signals, said stored binary signals representing a plurality of data field geometric configurations that are in approximate conformity with the geometric pattern of said presented character; sampling means coupled to said storage means for sampling successive areas of each of said data field geometric configurations to determine a correspondence between the information content of each sampled area and one of a predetermined number of binary characters and to provide an indication of such correspondence, each of said binary characters representing a pre-established area content; sequence detecting means coupled to said sampling means for registering, in sequence, each successive determined correspondence regardless of the particular data field geometric configuration the correspondence occurs in; and means for comparing the sequence of correspondence to a plurality of known sequences whereby said presented character is recognized as a known type when said sequence of correspondence is equal to that known sequence associated with said known type.

Claim 10

Character recognizing apparatus comprising: analyzing means for generating binary signals representing the information content of a geometric pattern corresponding to a presented character; storage means coupled to said analyzing means for storing said binary signals, said stored binary signals representing a plurality of data field geometric configurations that are in approximate conformity with the geometric pattern of said presented character; sampling means coupled to said storage means for sampling successive areas of each of said data field geometric configurations to determine a correspondence between the information content of each sampled area and one of a predetermined number of binary characters and to provide an indication of such correspondence, each of said binary characters representing a pre-established area content; sequence detecting means coupled to said sampling means for registering, in sequence, each successive determined correspondence regardless of the particular data field geometric configuration the correspondence occurs in; and means for comparing the sequence of correspondence to a plurality of known sequences whereby said presented character is recognized as a known type when said sequence of correspondence is equal to that known sequence associated with said known type.

Claim 11

Character recognizing apparatus in accordance with claim 7 wherein said storage means is comprised of a multiple of individual stages, each said stage adapted to store a binary signal; and wherein groups of said stages represent corresponding areas of said data field geometric configuration.

Claim 12

Character recognizing apparatus in accordance with claim 8 wherein said sampling means comprises: a plurality of coincidence means each including input terminals adapted to be supplied with the binary signals stored in selected stages of a group of stages, the number of coincidence means being equal to said predetermined number of binary characters, and each said coincidence means producing an output signal when a unique combination of binary signals is supplied to the input terminals thereof from a unique selection of stages; and means for coupling the input terminals of all of said plurality of coincidence means to those selected stages associated therewith in each succeeding group of stages in a sequential manner.

Claim 13

Character recognizing apparatus in accordance with claim 9 wherein said sequence detecting means comprises: a plurality of stepping means each including input terminals coupled to selected ones of said coincidence means, the number of stepping means being equal to the number of said known types, and each said stepping means including an output terminal adapted to be successively coupled to each input terminal thereof at a rate equal to the rate at whIch said input terminals of said coincidence means are coupled to succeeding groups of stages whereby an output signal produced by a coincidence means and supplied to an input terminal of said stepping means is provided at said output terminal if said output terminal is coupled to that input terminal; means for advancing the coupling of said output terminal of a stepping means from a given input terminal thereof to the next succeeding input terminal thereof in response to the provision of an output signal at the output terminal of said stepping means; and means for indicating when the output terminal of one of said stepping means has been advanced to be coupled to the last input terminal thereof, thereby detecting a completed predetermined sequence of produced output signals, whereby said presented character is recognized as the type associated with said one stepping means.

Claim 14

Character recognizing apparatus in accordance with claim 10 wherein said storage means includes: a plurality of columns of said stages to store corresponding columns of binary signals, said binary signals being shiftable through said storage means in a column by column time sequential manner, and wherein a column of binary signals represents the information content of a corresponding column of said geometric pattern; and means for detecting when columns of binary signals representing an entire data field geometric configuration have been shifted into said storage means and said data field geometric configuration is centered therein.

Claim 15

Character recognizing apparatus in accordance with claim 11 wherein said means for detecting comprises: sensing means coupled to a centrally disposed column in said storage means to sense the storage of a predetermined binary signal in any one of the stages in said centrally disposed column; and means responsive to said sensing means for generating at a predetermined time subsequent to the sensing of said predetermined binary signal a positioning signal indicating that said data field geometric configuration is centered in said storage means, said predetermined time corresponding to the time required to shift a column of binary signals through a determined number of columns.

Claim 16

Character recognizing apparatus in accordance with claim 12 wherein said means for advancing the coupling of said output terminal of a stepping means from a given input terminal thereof to the next succeeding input terminal comprises: counter means coupled to said output terminal for incrementing the count thereof in response to the provision of each output signal at said output terminal to thereby advance the coupling of said output terminal of said stepping means; and enable means coupled to said counter means and responsive to said positioning signal to enable said counter means to be incremented, said counter means being disabled in the absence of said positioning signal.

Claim 17

Character recognizing apparatus in accordance with claim 13 wherein at least one of said concidence means includes at least one input terminal coupled to a selected stage of a succeeding group of stages; at least one of said coincidence means includes at least one input terminal coupled to a selected stage of at least one preceding group of stages; and at least one of said coincidence means includes at least one input terminal coupled to a selected stage of a succeeding group of stages and at least one input terminal coupled to a selected stage of at least one preceding group of stages.

Claim 18

Character recognizing apparatus in accordance with claim 7 wherein said analyzing means comprises: a linear array of scanning means adapted for relative motion with respect to said presented character to serially produce voltages representative of the information content of each scanned section of said presented character; and quantizing means coupled to said scanning means for quantizing said serially produced voltages at plural quantizing levels to generate a corresponding Plurality of binary signals for each produced voltage whereby nonuniformities in said presented character are compensated.

Claim 19

Character recognizing apparatus in accordance with claim 15 wherein said storage means comprises a plurality of register means equal in number to said plural quantizing levels, each of said register means including a multiple of individual stages arranged in columns and rows and adapted to shift binary signals therethrough; one of said register means being adapted to receive binary signals generated by said quantizing means when said serially produced voltages exceed a first quantizing level to thereby store a normalized data field geometric configuration that represents a character formed of relatively thick strokes, another of said register means being adapted to receive binary signals generated by said quantizing means when said serially produced voltages exceed a second quantizing level less than said first quantizing level to thereby store a data field geometric configuration that represents a character formed of relatively normal strokes, and a further one of said register means being adapted to receive binary signals generated by said quantizing means when said serially produced voltages exceed a third quantizing level less than said second quantizing level to thereby store a normalized data field geometric configuration that represents a character formed of relatively thin strokes.

Claim 20

Character recognizing apparatus in accordance with claim 16 wherein said sampling means comprises: a plurality of coincidence means each including input terminals adapted to be supplied with the binary signals stored in selected ones of a row of stages for producing an output signal when a unique combination of binary signals are supplied from a unique selection of stages, the number of coincidence means being equal to said predetermined number of binary characters; and supply means for supplying in a row by row manner the binary signals stored in the selected stages of a row of each successive register means to the input terminals of all said coincidence means, said coincidence means being simultaneously supplied with binary signals from a register row.

Claim 21

Character recogniing apparatus in accordance with claim 17 wherein said supply means comprises: shift means coupled to each register means for advancing the contents of each row to the next preceding row in a timed periodic manner; and switch means coupled to each stage of a given row of each register means for simultaneously applying the input terminals of all said coincidence means to the selected stages in said given row, and for consecutively coupling all said coincidence means to each of said register means intermittent the period for advancing said contents of a row to a next preceding row.

Claim 22

Character recognizing apparatus in accordance with claim 18 wherein said sequence detecting means comprises: a plurality of stepping means equal in number to the number of said known types, each said stepping means including input terminals coupled to selected ones of said coincidence means and an output terminal adapted to be successively coupled to each said input terminal thereof, whereby an output signal produced by a coincidence means is proivided at said output terminal if said output terminal is coupled to that input terminal to which said coincidence means is coupled; means for advancing the coupling of said output terminal of a stepping means from a given input terminal to the next succeeding input terminal thereof in response to the provisions of an output signal at the output terminal of said stepping means; and means for indicating when the output terminal of one of said stepping means has been advanced to be coupled to the last input terminal thereof, thereby detecting a completed predetermined sequence of produced output signals regardless of the particular register means the selected combinations of binary signals resulting in said sequence of output signalS are stored in, whereby said presented character is recognized as the type associated with said one stepping means.

Claim 23

Character recognizing apparatus in accordance with claim 19 wherein said advancing means includes enabling means responsive to a signal indicating that an entire data field geometric configuraton has been shifted into said shift register means and said data field geometric configuration is centered therein for enabling said advancing means to couple said output terminal of a stepping means from one input terminal to the next succeeding input terminal.

Claim 24

Character recognizing apparatus in accordance with claim 20 wherein said storage means further includes positioning means for generating a positioning signal indicating that an entire data field geometric configuration has been shifted into said register means and said data field geometric configuration is centered therein, said positioning means comprising: sensing means coupled to a centrally disposed column in one of said register means to sense the storage of a predetermined binary signal in any one of the stages therein; and means responsive to said sensing means for generating said positioning signal at a predetermined time subsequent to the sensing of said predetermined binary signal, said predetermined time corresponding to the time required to shift a column of binary signals through a determined number of columns.

Claim 25

Character recognizing apparatus in accordance with claim 21 wherein said plurality of coincidence means includes: at least one coincidence means having at least one input terminal adapted to be supplied with a binary signal stored in a selected stage in a row succeeding said given row; at least one coincidence means having at least one input terminal adapted to be supplied with a binary signal stored in a selected stage in at least one row preceding said given row; and at least one coincidence means having at least one input terminal adapted to be supplied with a binary signal stored in a selected stage in a row succeeding said given row and having at least one input terminal adapted to be supplied with a binary signal stored in a selected stage in at least one row preceding said given row.

Claim 26

A character recognition system for identifying a presented character as being a known type, comprising: scanning means for generating an array of voltages representing the information content of a scanned pattern corresponding to said presented character; multi-level quantizing means coupled to said scanning means for quantizing said array of voltages at plural quantizing levels to form a corresponding plurality of weighted data field geometric configurations consisting of rows of binary signals; each of said weighted data field geometric configurations representing a normalized scanned pattern thereby compensating for nonuniformities in said presented character; means for correlating each row of binary signals of each weighted data field geometric configuration with predetermined binary words in a row by row sequential manner, a given row in each data field geometric configuration being correlated with said predetermined binary words before a next succeeding row is correlated; means for detecting each binary word with which each row is in correlation; means for registering the sequence of detection regardless of the particular weighted data field geometric configuration within which the correlated rows are included to determine the order in which said detected binary words are arranged to form a representation of said scanned pattern; and means for comparing said determined order with a plurality of pre-established orders, each said pre-established order being associated with a known type, whereby said presented character is identified as the known type associated with the one pre-established order with which said determined order favorably compares.

Claim 27

A character recognition system for identifying a presented character as being a known type, comprising: scanning means for generating an array of voltages representing the information content of a scanned pattern corresponding to said presented character; multi-level quantizing means coupled to said scanning means for quantizing said array of voltages at plural quantizing levels to form a corresponding plurality of weighted data field geometric configurations consisting of rows of binary signals; each of said weighted data field geometric configurations representing a normalized scanned pattern thereby compensating for nonuniformities in said presented character; means for correlating each row of binary signals of each weighted data field geometric configuration with predetermined binary words in a row by row sequential manner, a given row in each data field geometric configuration being correlated with said predetermined binary words before a next succeeding row is correlated; means for detecting each binary word with which each row is in correlation; means for registering the sequence of detection regardless of the particular weighted data field geometric configuration within which the correlated rows are included to determine the order in which said detected binary words are arranged to form a representation of said scanned pattern; and means for comparing said determined order with a plurality of pre-established orders, each said pre-established order being associated with a known type, whereby said presented character is identified as the known type associated with the one pre-established order with which said determined order favorably compares.

Claim 28

A character recognition system in accordance with claim 23 wherEin said correlating means comprises comparator means to which selected ones of said binary signals included in a row are supplied in a row by row manner by each weighted data field geometric configuration to determine if said selected ones of said binary signals form, in combination, one of said predetermined binary words.