Stanford Research Institute

Stanford Research Institute International

Daniel E Wolf

Hewitt D Crane

John S Ostrem

Samuel Lindenberg

In order to verify the authenticity of the signature of an individual, first a sample signature is written with a pen which can generate electrical signals representative of the forces exerted in the plane of the paper in the process of writing. These signals are sampled and then scaled so that the signature when displayed, can be fitted within a predetermined display area. The scaled signals generated from the sample signature are then stored. When it is desired to compare the sample signature with a signature which is subsequently written, called a specimen signature, the specimen signature is written and is processed for display in the same manner as was the sample signature. The scaled sample signature signals are then called out of storage and displayed, as are the scaled signals generated in the course of writing the specimen signature. Any deviations between the signatures are very readily detectable by observing the display.

Dynamic creation of signatures

When trains of signals, derived from a written template signature, are to be compared directly with trains of signals, derived from a written specimen signature, because no two written signatures are identical, there is a requirement for providing a certain degree of 'rubberiness' between the specimen and template signals. Previously, such rubberiness was achieved by breaking up the trains of signals into fixed portions and then conducting translation, stretch and contraction between these fixed portions in the course of the comparison. In this invention, advantage is taken of the existence of prominent landmarks in the trains of signals of the template and specimen signatures. The trains of signals are broken into segments by these landmarks. Segments of the specimen are stretched or contracted and translated to match corresponding segments of the template, to greatly simplify and enhance correlation process.

Signal train verification system using landmarks

Dynamic re-creation of signatures

When signals derived from a template signature are to be compared directly with signals derived from a specimen signature, because no two signatures are alike, there is a requirement for providing a certain degree of 'rubberiness' between the specimen and template signals. Previously such rubberiness was achieved by breaking the signal up into fixed portions and then conducting translation, stretch and contraction between these fixed portions in the course of the comparison. In this invention, advantage is taken of the existence of prominent landmarks in the template and specimen signatures, and the signatures are broken up between these landmarks for the purpose of achieving rubberiness and/or translation. The landmarks in the signals derived from the signatures to be compared are exactly aligned before the correlation process is undertaken.

Handwriting verification system using landmarks

Identification of a specimen signature is provided by comparing the specimen with a previously recorded template signature. The comparison is performed by recording, for both specimen and template signatures, signals representing forces in three directions while generating these signatures. For handwritten signatures these forces are called pressure, for the force perpendicular to the plane of the paper, X for the left-right force in the plane of the paper, and Y for the near-far force in the plane of the paper. The recorded specimen signals are divided into equal parts, for example halves, and successive comparisons are made of the equal parts, first without any displacement, then with relative displacements, then by expanding the length of parts relative to one another, then by reducing the length of parts relative to one another, to find the highest correlation values for all of these comparisons. These are combined and compared with a reference.

Handwriting verification system

A person, who is to have his identity verified, first writes his signature or any other appropriate group of characters or symbols several times with a special pen which produces signals representative of the writing forces in the plane of the paper and the writing pressure. From these a number of different parameters are derived. These parameters may be, for example, average value, energy, timing, number of zero crossings, etc. Average values and standard deviations are obtained for each of these parameters and these are stored as components of a template vector. In order to detect whether or not a later handwriting sample is authentic, a measure of the difference between the template vector and the later handwriting sample vector is calculated. The distinction between true signatures and forgeries is then made on the basis of this difference. If it is less than an appropriately selected value the signature is judged authentic while if it is above such value it is judged a forgery. The measure of difference may be in terms of RMS difference, for example, or say, average difference. Or, a signature may be deemed a forgery if any one component of the later written signature vector, or some combination of components, deviates beyond some threshold value.

Growing community of inventors

Menlo Park, CA, United States of America

Daniel E Wolf