Method for verifying a handwritten signature entered into a digitizer

• US 5,434,928 A
• Filed: 12/06/1993
• Issued: 07/18/1995
• Est. Priority Date: 12/06/1993
• Status: Expired due to Fees

• Current Assignee: NCR Corporation (NCR Voyix Corporation)
• Original Assignee: AT&T Global Information Solutions Company (NCR Voyix Corporation)
• Inventors: Moore, Clifton F. Jr., Wagner, Samuel J.
• Primary Examiner(s): Couso, Jose L.
• Application Number:
  US08/163,469
• Time in Patent Office: 589 Days
• Field of Search: 382/3, 382/13, 382/30, 382/34, 382/48, 902/6, 902/25, 235/380
• US Class Current: 382/187
• CPC Class Codes: G06V 40/30 Writer recognition; Reading and verifying signatures
• CPC Class Codes: G07C 9/247 electronically, e.g. by comparing signal of hand-writing with a reference signal from the pass

Abstract

A handwriting verification method which is less expensive to implement than previous methods, employs simpler algorithms than previous methods, and which requires fewer points than previous methods. The method includes the steps of determining a set of positional data points for a handwritten signature entered into the digitizer, determining a set of global variable values associated with the signature from the positional data points, determining head and tail strings having shape information for the signature, reading previously computed head and tail strings from a storage medium carried by the person, reading the verification information from the storage medium, determining a range of acceptance values for each global variable, and determining whether each of the global verification values lies within the range of acceptance. If at least one of the global verification values lies outside the range of acceptance, the method further includes the step of rejecting the subsequent signature as a forgery. If each of the global verification values lies within the range of acceptance, the method further includes the step of determining whether the head and tail strings of the subsequent signature equal the head and tail strings from the storage medium. If either of the head and tail strings of the subsequent signature are not equal to the head and tail strings from the storage medium, the method further includes the step of rejecting the subsequent signature as a forgery. Finally, if both the head and tail strings of the subsequent signature are equal to the head and tail strings from the storage medium, the method includes the step of accepting the signature as valid.

Claims

1. A method of verifying a handwritten signature entered into a digitizer comprising the steps of:

(a) providing a handwriting capture device including the digitizer and a hand-held stylus for capturing the handwritten signature;

(b) sensing the location of a series of signature points progressively traversed by the stylus during strokewise movement thereof by the digitizer for a plurality of signatures of a person;

(c) generating a stream of digital position data from the analog position data from the digitizer for the plurality of signatures indicating the location of the signature points on the surface of the digitizer, pen down time, and total time to sign;

(d) storing the position data in an ordered arrangement corresponding to the strokewise sequence of the data within the stream for the plurality of signatures;

(e) retrieving the position data from storage in a sequence related to the storage order thereof for the plurality of signatures;

(f) generating verification information including a count of the signatures, a sum of global variable values for the signatures, a sum of the squares of the global variable values for the signatures, and head and tail strings representing the average shape of the signatures by a processor coupled to the digitizer;

(g) storing the verification information in a storage medium carried by the person;

(h) determining a set of positional data points for a subsequent handwritten signature entered into the digitizer by a person;

(i) determining a set of global variable values associated with the subsequent signature from the positional data points by the processor;

(j) determining head and tail strings for the subsequent signature by the processor;

(k) reading the head and tail strings from the storage medium carried by the person by the processor;

(l) reading the verification information from the storage medium by the processor;

(m) determining a range of acceptance values for each global variable by the processor;

(n) determining whether each of the global verification values lies within the range of acceptance;

(o) if at least one of the global verification values lies outside the range of acceptance, rejecting the subsequent signature as a forgery;

(p) if each of the global verification values lies within the range of acceptance, determining whether the head and tail strings of the subsequent signature equal the head and tail strings from the storage medium;

(q) if both of the head and tail strings of the subsequent signature are not equal to the head and tail strings from the storage medium, rejecting the subsequent signature as a forgery;

(r) if either the head or the tail string of the subsequent signature are equal to the head or tail strings from the storage medium, accepting the signature as valid;

(s) compressing and packing the subsequent signature by the processor, including the substeps of

(s-1) retrieving the position data from storage in a sequence related to the storage order thereof;

(s-2) computing the locations of the signature points relative to a series of strokewise oriented guide lines by the processor, wherein each guide line has two end points, including a start point and a stop point;

(s-3) selecting only the position data corresponding to those of the signature points which do not fall within a rectangle about a corresponding guide line, the rectangle including a first line through the start point and perpendicular to the guide line, a second line through the stop point and perpendicular to the guide line, a third line on one side of the guide line and parallel to it, and a fourth line on the other side of the guide line and parallel to it;

(s-4) determining delta values for each guide line, each delta value being represented by a predetermined number of bits less than the number of bits representing the end points;

(s-5) discarding a predetermined number of lower bits from each delta value; and

(s-6) readjusting each delta value to limit error propagation; and

(t) storing the compressed-packed subsequent signature in another storage medium coupled to the processor.

2. The method as recited in claim 1, wherein step (j) comprises the substeps of:

(j-1) establishing a vector between each data point of the subsequent signature by the processor;

(j-2) assigning an integer direction value to each vector based upon the direction of the vectors;

(j-3) forming a data word from the series of integer values;

(j-4) representing a first series of like direction values from the beginning of the data word to a predetermined position in the data word by a single one of the like direction values to form the head string; and

(j-5) representing a second series of like direction values from a position after the predetermined position to the end of the data word by a single one of the like direction values to form the tail string.

3. The method as recited in claim 1, wherein step (m) comprises the substeps of:

(m-1) reading the sum of the global variable values, the sum of the squares of the global variable values, and the count of signatures from the storage medium;

(m-2) determining the mean MEAN[i] and standard deviation STD[i] of the global variable values;

(m-3) determining a minimum acceptance value using the formula

space="preserve" listing-type="equation">MIN[i]=MIN([MULT1[i]·

STD[i]-DECR[i], MULT2[i]·

MEAN[i]);

and

(m-4) determining a maximum acceptance value using the formula

space="preserve" listing-type="equation">MAX[i]=MAX(MULT3[i]·

STD[i]+INCR[i], MULT4[i]·

MEAN[i]).

4. The method as recited in claim 1, wherein step (t) comprises the substeps of:

(t-1) storing one of the end points and a corresponding readjusted delta value for each guide line.

5. The method as recited in claim 1, wherein the storage medium carried by the person comprises a SMART card.