Complex pattern recognition method and system

Claim 2

In an apparatus for identifying patterns by means of a pattern registration means and a data processing means, wherein each pattern comprises at least one stroke based on relative movement between a pattern-forming means and a pattern-accommodating means of said stroke registration means, said stroke registration means being operative to generate first signal information represented by first signals specifying relative position along a stroke, said data processing means being provided with prestored second signals and being operative to generate third signals representing third signal information, said third signal information comprising specification of identity labels for strokes for use in developing coded information specifying said pattern, characterized in that: said data processing means includes means for designating from said first signal information a set of first data points, which may include points selected from or interpolated on the basis of said first signal information, along said stroke, wherein said second signal information defines templates, each said template comprising a set of second data points defining relative position values and set of permissible tolerance values for second data points, at least one of said permissible tolerance values differing from at least one other of said permissible tolerance values; and said data processing means further includes means for comparing said first data points with said second data points to determine which of said templates provides best correspondence between said first and second signal information to produce said third signal information, whereby pattern identification accuracy is enhanced and ambiguity is reduced by selective disregard of information about stroke characteristics.

Claim 3

In the apparatus according to claim 1, characterized in that said pattern-accommodating means is a signal-receiving means responsive to relative movement of said pattern-forming means.

Claim 4

In the apparatus according to claim 1 or 2 wherein said pattern-forming means comprises a writing implement and said pattern-accommodating means comprises a pressure-sensitive position-registration tablet.

Claim 5

In the apparatus according to claim 1 or 2 wherein said pattern-forming means and said pattern-accommodating means are characterized by an instrumented writing implement operative to generate transducer signals specifying pressure and direction versus time in response to an act of writing with said implement on a friction surface.

Claim 6

In the apparatus according to claim 1, 2, 3, or 4, characterized in that said data processing means includes means for rotating said first signal information relative to position values of said second signal information.

Claim 7

In the apparatus according to claim 1, 2, 3, 4, or 5, characterized in that said data processing means includes means for displacing said first signal information relative to position values of said second signal information.

Claim 8

In the apparatus according to claim 1, 2, 3, 4, 5, or 6, characterized in that said data processing means includes means for preserving selected absolute stroke position values of said first original information independent of said identifying step and to employ said selected absolute stroke position values to resolve ambiguity among recognized patterns only in the event a plurality of patterns is indicated for a signal accumulation of identity labels.

Claim 9

In a method employing an apparatus according to claim 1, said method for identifying patterns by means of a pattern registration means and a data processing means, wherein each pattern comprises at least one stroke based upon relative movement between a pattern-forming means and a pattern-accommodating means of said registration means, said registration means being operative to generate first signal information represented by first signals specifying relative position along a stroke, said data processing means being provided with prestored second signal information represented by second signals and being operative to generate third signals representing third signal information, said third signal information comprising specification of identity labels for strokes for use in developing coded information specifying said pattern, characterized in that for at least one stroke: said data processing means designates from said first signal information a set of first data points, which may include points selected from or interpolated on the basis of said first signal information, along said stroke, wherein said second signal information defines templates, each said template comprising a set of second data points defining relative position values and a set of permissible tolerance values for second data points, at least one of said permissible tolerance values differing from at least one other of said permissible tolerance values; and said data processing means compares said first data points with said second data points to determine which of said templates provides best correspondence between said first and second signal information to produce said third signal information, whereby pattern identification accuracy is enhanced and ambiguity is reduced by selective disregard of information about stroke characteristics.

Claim 10

In the method according to claim 8, characterized in that said permissible tolerance values are deviation values which differ as between at least two of said second data points for a single template.

Claim 11

In the method according to claim 8 or 9, characterized in that, in the absence of acceptance of an identity label, said data processing means modifies said first signal information by adding or subtracting an incremental numerical value of at least one of said first data points and again compares said first data points with each said template.

Claim 12

In the method according to claim 10, characterized in that said modifying by data processing means comprises rotating said first signal information of at least one of said first data points relative to said second data points.

Claim 13

In the method according to claim 8, 9, 10, or 11, characterized in that, in the absence of acceptance of an identity label, said data processing means modifies said first signal information to effect a displacement of said first data points relative to said second data points and again compares said first data points with each said template.

Claim 14

In the method according to claim 12, the improvement characterized in that said modifying by said data processing means comprises shifting said first data points relative to said second data points along stroke length.

Claim 15

In the method according to claim 8, 9, 10, 11, 12 or 13, the improvement characterized in that said data processing means resolves said first signal information with an angular resolution at each one of said first data points to within no greater than 1.5 degrees.

Claim 16

In the method according to claim 8, 9, 10, 11, 12 or 13, characterized in that said data processing means resolves said first signal information with an angular resolution at each one of said first data points to within no greater than 0.15 degrees.

Claim 17

In the method according to claim 8, 9, 10, 11, 12, 13, 14 or 15, characterized in that said data processing means compares said first data points with said second data points by computing the closeness between each one of said first data points and said second data points to obtain a distance metric value.

Claim 18

In the method according to claim 8, 9, 10, 11, 12, 13, 14 or 15, characterized in that said data processing means calculates distance between corresponding first data points and second data points to obtain difference values; divides each one of said difference values by the corresponding permissible angular deviation value to obtain quotient values; and sums all quotient values to obtain a distance-metric value.

Claim 19

In the method according to claim 17, characterized in that said data processing means squares each one of said quotient values prior to summing.

Claim 20

In the method according to claim 16, characterized in that said data processing means compares said distance metric value for each said template with an upper decision threshold value, with a lower decision threshold value and with a difference decision threshold value in order to obtain said identity label.

Claim 21

In the method according to claim 19, characterized in that said data processing means assigns a specific identity label of one of said predefined template strokes if the least distance metric value is less than the lower decision threshold value and if the next least distance metric value is greater than the upper decision threshold value.

Claim 22

In the method according to claim 19, characterized in that said data processing means assigns a specific identity label of one of said predefined template strokes if the least distance metric value is less than the upper decision threshold value and if the next least distance metric value is greater than the sum of the least distance metric value and the difference decision threshold value.

Claim 23

In the method according to claims 8 through 21, characterized in that: said third signals representing said identity label are combined with other third signals representing other identity labels so obtained to form fourth signals -representing a group of related identity labels; and said fourth signals are compared against fifth signals, said fifth signals representing an index to character labels specified by sets of related identity labels to produce sixth signals, said sixth signals representing a character label corresponding to said group.

Claim 24

In the method according to claim 22, characterized in that said sixth signals are employed to form a visual pattern representing the character identified by said character label.

Claim 25

In the method according to claim 22, characterized in that said data processing means uses information regarding sequence of said identity labels to produce said sixth signals representing character labels.

Claim 26

In the method according to claims 8 through 24, characterized in that said data processing means compares said second signal information with said first signal information independent of absolute stroke position values.

Claim 27

In the method according to claim 25, characterized in that said data processing means preserves selected absolute stroke position values of said first signal information independent of said identifying step, and employs said selected absolute stroke position values to resolve ambiguity among recognized patterns only in the event a plurality patterns is indicated for a single accumulation of identity labels.