CA2105926A1 - Fingerprint verification system - Google Patents
Fingerprint verification systemInfo
- Publication number
- CA2105926A1 CA2105926A1 CA002105926A CA2105926A CA2105926A1 CA 2105926 A1 CA2105926 A1 CA 2105926A1 CA 002105926 A CA002105926 A CA 002105926A CA 2105926 A CA2105926 A CA 2105926A CA 2105926 A1 CA2105926 A1 CA 2105926A1
- Authority
- CA
- Canada
- Prior art keywords
- light modulator
- spatial light
- fingerprint
- correlation filter
- beam path
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
- 238000012795 verification Methods 0.000 title claims abstract description 62
- 230000003287 optical effect Effects 0.000 claims description 57
- 238000000034 method Methods 0.000 claims description 15
- 230000002596 correlated effect Effects 0.000 abstract description 3
- 238000012545 processing Methods 0.000 description 9
- 238000004422 calculation algorithm Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 7
- 238000013459 approach Methods 0.000 description 5
- 238000013461 design Methods 0.000 description 5
- 238000003860 storage Methods 0.000 description 5
- 230000001427 coherent effect Effects 0.000 description 4
- 238000013500 data storage Methods 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 4
- 238000003909 pattern recognition Methods 0.000 description 4
- 238000007781 pre-processing Methods 0.000 description 4
- 230000002123 temporal effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 239000002131 composite material Substances 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000036039 immunity Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 101100508878 Escherichia coli (strain K12) rsfS gene Proteins 0.000 description 1
- 101100292616 Saccharomyces cerevisiae (strain ATCC 204508 / S288c) SLM3 gene Proteins 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 235000013361 beverage Nutrition 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000003750 conditioning effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000839 emulsion Substances 0.000 description 1
- 230000008014 freezing Effects 0.000 description 1
- 238000007710 freezing Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000001093 holography Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000010348 incorporation Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000005693 optoelectronics Effects 0.000 description 1
- 230000010287 polarization Effects 0.000 description 1
- 238000002310 reflectometry Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000009885 systemic effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C9/00—Individual registration on entry or exit
- G07C9/20—Individual registration on entry or exit involving the use of a pass
- G07C9/22—Individual registration on entry or exit involving the use of a pass in combination with an identity check of the pass holder
- G07C9/25—Individual registration on entry or exit involving the use of a pass in combination with an identity check of the pass holder using biometric data, e.g. fingerprints, iris scans or voice recognition
- G07C9/257—Individual registration on entry or exit involving the use of a pass in combination with an identity check of the pass holder using biometric data, e.g. fingerprints, iris scans or voice recognition electronically
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1365—Matching; Classification
Landscapes
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Collating Specific Patterns (AREA)
- Holo Graphy (AREA)
Abstract
FINGERPRINT VERIFICATION SYSTEM
Abstract A fingerprint verification system is disclosed which can be used as a fast and effective security device to verify that an applicant to a service protected by the system is authorized to use the system. This security check is performed through a correlation operation. Electronic information allowing the creation of a correlation filter on a spatial light modulator (SLM) is input from a smart card or fingerprint database to an SLM. The fingerprint or cluster of fingerprints of the applicant are then correlated through the SLM in order to derive a comparison with those electronically stored. Although the system used electronically stored pattern information, the correlation is performed optically.
Abstract A fingerprint verification system is disclosed which can be used as a fast and effective security device to verify that an applicant to a service protected by the system is authorized to use the system. This security check is performed through a correlation operation. Electronic information allowing the creation of a correlation filter on a spatial light modulator (SLM) is input from a smart card or fingerprint database to an SLM. The fingerprint or cluster of fingerprints of the applicant are then correlated through the SLM in order to derive a comparison with those electronically stored. Although the system used electronically stored pattern information, the correlation is performed optically.
Description
2iO5926 FINGERPRINT VERIFICATION SYSTEM
Backqround of the Invention 1. Field of the Invention This invention relates to a fingerprint verification device.
Backqround of the Invention 1. Field of the Invention This invention relates to a fingerprint verification device.
2. Description of Related Art The roots of optical processing go back to the last century when Ernst Abbe performed some pioneering research.
Coherent optical pattern recognition has been studied since its introduction by Anthony Vander Lugt in 1964 (see Vander Lugt, A., Siqnal Detection bY ComPlex SPatial Filterinq IEEE Transactions on Information Theory, vol.IT-10 no. 2 p.139, Apr 1964). This important step wa~ made possible by the development of both a power~ul coherent light source (the laser, in 1961) and off axis holography, by Leith and Upatnieks, also in 1964. From this time to the present, optical techniques for pattern recognition have been known to be extremely powerful because of their speed, high discrimination powers, and noise immunity.
The fundamental phenomenon upon which optical correlators rely is the natural ability of a lens to extract a two dimensional Fourier transform of an input signal. The speed at which the lens will perform the transformation is independent of the input signal or object; a complex input is processed ~ust as fast as a simple one. The computation is done completely in parallel, and occurs literally at the speed of light. This speed represents the most significant advantage that optical processors have over digital devices, and is the reason why so much research has been conducted in this field. Optical devices based on the transformation properties of a lens are commonly referred to as Fourier optical signal processors.
Unfortunately, optical correlation devices suffer from some severe practical difficulties which have prevented them from becoming useful in the real world. In order to exploit the processing speed of the optical technique, input data must be introduced at an extremely high rate, and of course the output must be removed just as fast. Furthermore, in order to compare one pattern with another, a reference or target pattern must be input as well. This reference data is input in the form of a correlation filter, and must be physically located at the transform plane of the correlator. These two input requirements nece~sitate fast, high resolution devices which can compete with the ~nformation bandwidth capabilities of modern digital hardware.
This data transfer problem has always presented a signiflcant bottleneck in Fourier optical signal processing device~. However, this problem is not relevant in applications where high speed, high volume data processing is not a necessity.
Fingerprint verification iq ~ust such an application. In fingerprint verification a device reads recorded information about a reference fingerprint pattern and usQs this information to determine or verify that the reference pattern and an input pattern belong to the same individual. There is no requirement to compare a fingerprint with a large number of reference ~ - .- ~ . , - ... .. ,, .. . ~ , ~105926 patterns, nor is there any need to compare many input patterns with a single reference. A single input pattern must be compared with a single reference pattern, or possibly with a small number of reference patterns. ~revious attempts at optical fingerprint S pattern verification using Fourier techniques have made use of holographic matched filter (HMF) as the correlation filter. The HMF is typically processed on a card which is carried by the user.
However, a disadvantage of optical processing systems which does affect fingerprint verification utilising a Hl~F as the correlation filter is that the position and orientation of the correlation filter is extremely critical to the correct operation of the system. This positional tolerance is on the order of microns, and the rotational orientation must be within one or two d0grees of the optimal position. The mechanical positioning problem i5 therefore extremely difficult, requiring high resolution devices and extremely accurate alignment techniques.
Further, a serious systemic disadvantage is the inflexibility imposed by the holographic matched filter. The necessity of having a holographic matched filter precludes the possibility of transferring filter data by any means other than the physical transport of the original or its copies (which must be fabricated optically as well). The matched filter must be physically present in the frequency plane of the correlator, and there is currently no electronic device which will provide the resolution required to represent a holographic matched filter.
This invention seeks to overcome drawbacks of known ., , , :. . ~ , , : . , , , -... . ~ ,. . ~ . . ; , ,.
fingerprint pattern recognition devices.
SummarY of the Invention According to the present invention, there is provided a fingerprint verification device comprising the following: a source of light for providing an illuminating be~m along a beam path; input means located in the beam path including means to receive at least one fingerprint of an individual (either by the individual placing his finger(s) on the input means or by placing a fingerprint(s) recording in the input means) for modulating the ~eam with data from at least one fingerprint to form an information beam; optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plans; a spatial light modulator fixed in said transform plane; a card reader for reading encoded data from a card representing a correlation filter to be written to said spatial light modulator; means for receiving ~aid encoded data from said card reader and for writing a correlation filter to said spatial light modulator in accordance with said encoded data; verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between the Fourier transform of a reference fingerprint(s) which is derived from filter information contained on a card and written to the spatial light modulator, and the Fourier transform of the fingerprint(s) input.
According to another aspect of the invention, there is provided a method of generating a correlation filter for the . s . . . . -, , :. . . -fingerprint verification device comprising the steps of: forming a correlation filter from each of said at least two fingerprints of an individual which said input means is designed to receive;
superposing each said formed correlation filter to record a correlation filter.
According to another aspect of the invention, there is provided a method of generating a correlation filter for the fingerprint verification device comprising the steps of:
recording a correlation filter from a cluster of all of said at least two fingerprints of an individual which said input means is designed to receive.
Brief Description of the Drawin~s In the drawings which illustrate example embodiments of the invention:
figure 1 i9 a schematic diagram of a verification device made in accordance with this invention; and figure 2a, 2b and 2c are schematic views which illustrate alternate optical corxelation arrangements.
Detailed Description of the Preferred Embodiments In overview, recognising that the filter positioning problem could be solved by permanently fixing and aligning the filter in the proper position at the transform plane, a spatial light modulator ~SLM) is introduced into the transform plane and is fastened there at the time of manufacture. Correlation , 25 filters can then be electronically written onto the SLM, avoiding : . . , , , . ., -the need to mechanically position and align a filter for every correlation operation.
An electronically addressable filter in a correlator allows filter design and calculation to be carried out electronically with a computer. This fact alone makes the verification system far superior to one which utilizes a standard holographic matched filter. Computer generated filters need not be recorded using holographic techniques, which require sophisticated equipment such as lasers and vibration controlled environments. ~olographic matched filters require a holographic--capable recording medium, and finally a stable filter medium which when fixed onto a card to be carried by the user must be able to withstand the rigours of wallets and purses, hot sunlight, freezing temperatures, and spilt coffee and alcohollc beverages. Storlng the filter in a digitally encoded form permits the use of cards carrying encoded data, such as "smart cards", or central databases for filter data storage and retrieval.
Computer generated filters can be far more flexible ln the types of operations which can be performed on an input.
These filter algorithms can be very difficult if not practically impossible to implement optically (holographically). The various filter implementations can optimize the behaviour of the correlator with respect to noise performance, distortion invariance, and discrimination sensitivity. This is accomplished through careful analysis of the recognition problem to be solved.
The filter algorithm can then be tuned to the particular input, , . ~ . ~ ' ' ' . ! . , 2~5926 conditions, and application.
Fingerprint verification is extremely well suited to this application's specific tunability of the computer generated filter. The device has only to compare two sets of fingerprint patterns, and will not be required to differentiate between other types of objects. This allows the filters to be optimized to operate on specific characteristic features of fingerprint patterns which will provide the maximum amount of discrimination.
Distortion and noise characteristics can be studied and accounted for in the filter.
The introduction of an SLM upon which a correlation filter is written opens up a far broader range of useful applications, and makes the system realizable with current technology. The electronic filter data xepresentation permits easy communication of the data from one location to another, and eliminate the severe problems a~sociated with recording, manufacturing, alignment, and reliability of holographic matched filters.
The verification is performed through a hybrid optical/electronic correlation device. The actual correlation is performed optically, while the correlation signal is analyzed electronically. The ability to read stored reference pattern information is made possible through the introduction of a spatial light modulator (SLM3 at the filter plane of the optical correlator.
In practice, with reference to figure 1, a subscriber (an authorized user of a service which utilizes the fingerprint . . . : , . .
. ~ ...: . . . : . :
- . ; . . . .
. . ., . . , :.
2~05926 verification technology) inserts her card 10 into the fingerprint verification device (the verifier) 20, enabling the encoded fingerprint filter information to be txansferred by a reader 22 from the card to the processor 24. This unique information is used by the processor to create an optical filter which is transferred onto the SLM 26, which is contained within the verification device 20. The subscriber then places her fingers on the input screen 28 of the verifier, and the system performs an optical correlation operation, which compares the subscriber's fingerprint(s) with those represented by the encoded fingerprint information stored on the subscriber's card. The output from the SLM is an optical signal whose characteristics depend on the degree of correlation between the input and reference ingerprints. This optical signal, after passing through lens 34 and onto the optical detector 36 where its intensity digtribution i9 detected by the optical detector 36, is analyzed by the processor 24 which decides whether the output signal represent~ a sufficient correlation to indicate a match between the input and reference signals. A positive verification results in the processor 24 allowing the subscriber access to the desired service.
The verification procedure will occur very quickly, on the order of one second or less. The time will be primarily taken up by the various electronic systems which make up the device. The optical correlation operation will occur in several nanoseconds (several billionths of a second), the time required for light to travel from the input to the output plane. The i g verification operation should take no more time than is currently required by systems requiring users to enter PINs (personal identification numbers). This makes the device practical and acceptable as a commercial security system for consumer applications such as automated bank teller (ATM) machines and point of sale credit and debit card machines, while providing tremendously improved security against unauthorized use of such services. The high discrimination capability of Fourier optical pattern recognition techniques will ensure accurate verification.
The encoded representation of the filter information enables the incorporation of a card as an integral component of the security system. Encoded information on a card provides a reliable and readable storage medium for the reference pattern information. The same card can also carry information relevant to the services and privilege~ which the particular card entitles the holder.
The pattern verification is carried out by an optical correlator indicated generally at 38 and which comprises lenses 32 and 34, along with the SLM 26 and a detector 36.
Alternatively, the lenses could be replaced by holographic optical elements which also perform a Fourier transform. The method by which the fingerprint pattern is introduced to the correlator 38 is dependent on the input system, indicated generally at 29. The function of the input system is to generate a high contrast, noise and distortion free image of a live fingerprint(s) pattern and deliver it to the input of the optical correlator 38. The input system 29 may be an all--optical system ...... , . ~, .. . . .. .... .. . . . . . ...... . .. . .
.
.. .. .. . . ..
1 o or a hybrid opto--electronic system which would consist of image capture and display devices. The electronic system has the added capabilities of introducing previously recorded and stored input patterns as well as preprocessing of the input pattern. As illustrated in figure 1, the input system 29 comprises one or more lasers 30 along with an expander lens 31 and a collimator lens 33 to illuminate a prism 35 with a beam 37. One face of the prism forms the input screen 28. In an alternative input system, laser 30 may be replaced with an incoherent light source.
An optical correlator requires some degree of coherence in the optical signal as it passes through the system. Coherence i5 described in terms of spatial and temporal characteristics, and assumes a significant degree of polarization of the light.
Spatial coherence describe~ how light at one point on a plane perpendicular to the optical axis correlates in phase to light at another point on the plane. Temporal coherence relates the correlation of light at one point in space but at different points in time. If a broad band optica~ source is utilized, then ~patial and temporal coherence can be controlled by using source encoding and/or diffraction gratings respectively. Multichannel correlators use a diffraction grating to spatially isolate each wavelength at the filter plane. An alternative to a white light source is a series of lasers, each with a unique wavelength, which provides the advantage of higher spectral power. The SL~
based correlator described here will function with any of the possible light sources which are acceptable for Fourier optical processing devices.
;
~ .
210~926 An all--optical input system will utilize the principle of total internal reflection (~IR) to read the pattern formed by the furrows of the input fingerprint pattern. A furrow will create an air space over the surface of a glass screen, allowing 5light which is internally reflecting from the interior surface of the screen to proceed unimpeded. Ridges, however, will be in contact with the surface, where they will scatter and absorb a portion of the illuminating light. This effect is known as frustrated total internal reflection (FTIR). When the TIR light 10signal is observed, the primary signal will be due to the furrow pattern.
If the scattered light from the ridge pattern is observed, there may be significant advantages in terms of noise immunity from latent fingerprints. The scattered light signal 15will be weaker than the TIR signal from the furrows. Despite thls cost in illumlnation efficiency, the improved signal to noise ratio (SNR) of the FTIR signal from the ridges will most likely result in improved performance of the verification system in a real world environment.
20For both of the preceding all--optical input schemes, the light signal from the input pattern enters the optical correlator directly, without further interference or signal conditioning from any active electronic devices.
The alternative to the all--optical input system is an 25electronlc imaging and display system which captures the image of the input pattern and displays it on a ~evice such as an SLM.
~: The image capture may be performed by an optical array detector, : ~ :
210~926 scanner or similar device. The image can then be displayed at the input of the optical correlator 38 by illuminating the display device, either in transmission or reflection. This scheme permits preprocessing of the input pattern. PreproceSsing can provide added flexibility and power to an optical processor.
Alternatively, an optically addressable SLM can be used at the correlator input. This type of device allows a spatial pattern of light on one side of the SLM to affect the reflectivity on the other. Optically addressable SLMs are commonly used in optical signal processing devices to perform incoherent to coherent light conversion.
Input pattern illumination for the electronic imaging and display input may be the same as for the all--optical scheme;
TIR or FTIR could be exploited to create a high contrast image of the pattern.
The optical signal carrying the fingerprint pattern information proceeds into the correlator 38 and is transformed by the Fourier transform lens 32. Such a lens will perform an optical transform of the input pattern and display the transform at the Fourier transform plane of the correlator. The Fourier transform plane is typically located at the focal plane of the transform lens 32. The SLM 26 is positioned at the transform plane, and the optical transform of the input pattern passes through the SLM. A correlation filter is written onto the SLM, modulating the optical transform through a multiplicative operation with the function represented on the SLM. The second Fourier transform lens 34 i5 positioned in such a way that the . -210S92~
output of the SLM is re--transformed, resulting in the display of a correlation signal at the output plane of the correlator.
The correlation signal represents the transform of the product between the correlation filter written to the spatial light modulator - i.e., the reference fingerprint represented by the data read from the card 10 - and the Fourier transform of the fingerprints of the subscriber at the input screen 28. The correlation signal is usually accompanied by DC, convolution, and noise signals. These additional signals and their relative strengths depend on the correlator architecture and the filter characteristics.
The actual function represented on the SLM for correlation with a reference pattern can be arrived at through numerous method~, and they are described by various names lndicating their particular attributes. This flexibility is in fact a tremendous advantage of electronic filter calculation.
Whatever the algorithm, all the filters are designed to produce a corr~lation signal at the output of the correlator.
An array detector 36 at the output plane of the correlator intercepts the correlation signal, converting it to its electronic analog. This electronic signal i~ then captured, stored and analyzed by the processor 24 (electronic computer).
Through analysis of specific characteristics of the correlation ; signal the processor decides whether or not there is sufficient correlation between the input and reference patterns to warrant - a positive verification. The decision of the processor may be signalled to a host system, which then takes the appropriate : . - . -:. .
. . ~ . - - ,: -action.
The SLM is a device which can spatially m~dulate (in two dimensions) the amplitude and/or phase of light in either transmission or reflection. The spatial resolution of an SLM
varies widely and depends on the type of SLM. SLMs can be divided into two main categories, optically addressable and electrically addressable. In the optically addressable SLM
(OASLM) a pattern of light (coherence is irrelevant) is written onto the device, imposing an associated spatial modulation of a second (coherent) beam of light. In the electrically addressable SLM (EASLM) the magnitude of modulation of individual cells or pixels are controlled electrically, with each pixel having its own unique address.
O particular intere~t is the liquid crystal television or dlsplay (LCTV), These devices are widely used as flat panel displays for portable computers and calculators, control panel displays on equipment and instrumentation, and as video and pro~ection display devices for small televisions and projectors.
LCTVs are relatively inexpensive, have satisfactory spatial resolution, and can be made to modulate light in amplitude and/or phase. LCT~s have proven themselves to be mas6 producible and reliable, and have minimal power requirements.
Current technology does not provide the optical system designer with an SLM possessing the spatial resolution capabilities of a holographic recording emulsion, but the resolution is sufficient for correlation operations in an optical data processing device. However, the low spatial resolution - - .
.,, . , ., . . , . .. . ... .- .. - .. ~
210~926 results in a low diffraction efficiency, while the information content required to implement a classical matched filter, even with a 128 x 128 pixel S1M, i8 not small. Advances in filter algorithms has resulted in superior approaches to filter implementations which do not simply attempt to implement a low resolution holographic matched filter.
It must be noted that a holographic matched filter employs a carrier frequency (created by the reference beam) to encode the complex amplitude information on an amplitude only format. This encoding scheme stores both amplitude and phase information, which naturally requires a high information storage capacity from the medium, i.e. high resolution. If the filter modulates only the phase of the light, not only is the efficiency improved (there are no absorption losses in this case), but the information content ls sub3tantially reduced. Such a filter is called a phase only filter (POF), and hag been shown to perform a~ well or even better than a matched filter.
Even further ~aving~ in data storage can be had if the filter function i~ binarlzed. This proce~s conver~s a continuous pha~e function lnto a two level function. There are once again numerous algorithms for binarization, but the data storage requirements can be seen to be extremely minimal. Such a filter is called a binary phase only filter (BPOF), and has also been shown to perform extremely well in spite of the reduced information content.
A 3ignificant advantage of the SLM as a filter is the flexibility afforded by its ability to rapidly change filters ,,, ", . ".,~ . "' 1 " ,, , ,, ,, . ~,, ., . . , :, " :, ", ' ,' ., " , . . .
. . . :. :, ,. ~
- ~lD~926 without mechanically replacing the filter. An SLN can therefore be incorporated into the correlator and aligned at the time of manufacture, avoiding the difficulty of mechanically aligning a HMF in the transform plane at every verification operation.
Holographic filter alignment is known to be very critical for correct operation of a HMF correlator, requiring precision on the order of microns. This places a tremendously high emphasis on the mechanical accuracy and reliability of a positioning system.
The operation of the SLM can be in amplitude, phase, or full complex (amplitude and phase) modulation. The modulation can also be in binary form, as opposed to continuous modulation.
Binary modulation, as previously stated, has the advantage of greatly reduced data storage requirements.
~nother advantage of the SLM--based correlator is that the design and fabrication of the filter can be performed electronically, introducing far more flexibility in filter design. A computer generated filter (CGF) can implement operations which can be very difficult or practically impossible to perform with an optically recorded filter. This make~ filter design and testing much faster and more effective. For example, the SLM can compen~ate for any misalignment in the fingerprint on the input screen of the verifier. This may be accomplished by electronically rotating the ~ilter on the fixed SLM. The actual degree of misalignment does not need to be known since the filter on the SLM may be electronicàlly rotated until the maximum output signal is found. Alternatively, the filter on the SLM
could be made rotationally invariant. Furthermore, the pattern :., , .. , , ., ........... . . - .. .. :.... . ......... .
,, ; . , . . . . - ~ , .
- 17 _ on the filter may be electronically made larger or smaller to accommodate scale changes of the input fingerprint. Modern digital techniques permit preprocessing of the reference fingerprint pattern to optimize the correlators performance in terms of expected noise and input distortions. Significant advances in filter algorithms provide the system designer with a large number of starting points from which to start developing an application specific filter.
Electronic or digital storage of the reference fingerprint filter information avoids the practical difficulties associated with holographic recordings and recording media which must be dealt with in a HMF--based sy~tem. Instead of the holographic recording medium on the card, the information required to write the filter onto the SLM is stored on a card such as a smart card. The verifier reads this information from an inserted card and writes the pre~cribed correlation filter onto the SLM.
Encoded fingerprint filter information storage represent~ an enormous advance over the HMF in the collection, processing and distribution of fingerprint data. The enrolment procedure can now be an easily di~tributed operation.
Fingerprint scanners will simply digitize subscribers' fingerprints and relay the information to a filter-generating computer. The ease with which digital information can be transmitted makes the actual location of the computer irrelevant.
Moreover, advances in computer and communications technologies make the po~sibilities of a near real time enrolment a realistic 21~S926 possibility. With such a system a new subscriber would insert his newly issued blank card into the enrolment device, apply his fingerprints to the screen, and remove the initialized card ~everal seconds later, thus completing the enrolment. The use of two or more fingerprints (a cluster) offers significant advantages over the use of just a single digit. Two approaches are possible. The first involves the digitization of the cluster as a single image, from which a filter is generated just as it would be generated for a single digit. (Digitization and filter generation are independent of the image content.) There is obviously more identifying information in a cluster than in a single digit, resulting in a filter which has the same amount of information ~the same number of data bits) but with more unlque information content. This information now includes relative finger position (i.e. relative length and orientation) because of the ~patial relationships between the features of one digit of the cluster with the all the features on the other digit(g) .
A second approach to the use of a cluster is to digitize each finger independently. Independent filters could be generated for each finger in the cluster, or a single composite filter could be derived. The case of two or more independent filters clearly represents more information than a ~ingle filter. At verification, the cluster is applied to the input of the verification device which in this application uses one or more sources of light. The card, holding independent correlation filter data for each digit in the cluster, is , ,' ' ~ .. ~ " ~' , ''.,' ,', ':' ' ' :: . . , : . . . .
-: . : . -210592~
inserted into the card reader as usual. The verification device then displays each filter sequentially, verifying the presence of each digit in the cluster. A unique feature of this approach is that the correlation peak location at the output is dependent on the actual location of the target digit on the input screen.
By tracking the correlation peak location for each filter, the cluster shape (i.e. the relative finger lengths and widths) can be reconstructed. This provides additional biometric information to the system which can be used to improve system performance beyond what would be possible with only a single digit.
Alternatively, a single composite filter can be constructed from the digitized images of each finger in the cluster. This will provide a single filter capable of correlating with each digit in the cluster, and also provide relative digit location in the lnput plane of the verification device.
Encoded filter data also provides a convenient format for information storage and retrieval. The verification technology then becomes applicable where ever a centrally accessible database of authorized subscribers is desirable. This type of security system would not require the ~ubscriber to carry a filter card at all. For example, the verifier could access the subscriber's encoded fingerprint information from a central data~ase by cross referencing with the user's P.I.N. entered on a keyboard. The fingerprint of the user is then correlated using~
a filter written onto the SLM with the centrally stored pattern information.
If the database is small enough and the S~M fast . :. : . : . ..... .. ~ . , :
, ' , .. !.. : . . ' . , , ~ . . . . .
: ' ,, .. . '' ~ . ' ' , ' '. ', , ' ' ,' ': ' '' . '' , '' ' . ' .' . ~ ,.
enough, a match for an input fingerprint can be searched for by sequentially scanning through the database. This approach may be desirable for access to high security installations or systems, etc., where there are relatively few authorized per30nnel. Current LCTV devices have scan rates of 60 Hz., allowing a search through a thirty record database in half a second. The ~peed can be expected to improve with advances in technology.
A further advantage of the digital filter format is the flexibility it offers for upgrading and improving the system.
Improved filter designs are readily implemented by the filter generating computer. Existing cards can be upgraded at the subscriber's convenience at any enrolment facility. Upgrading of S~M drivers and electronics in the verifiers can be tolerated beaause o~ the electronic nature of these devlces. Advances in relevant technology can be exploited, yet backwards compatibility can be always be ensured through software.
It should be noted that the optical arrangement for the correlator 38 illustrated in figure 1 is by way of example only.
Other optical arrangements are known. In this regard, reference is made to figures 2a, 2b, and 2c which shows a few of the known alternate arrangements for an optical correlator with P1 indicating the input, P2 the correlation filter, and P3 the detector. It will be recognised that figure 2b illustrate~ the optical arrangement for the correlator illustrated in figure 1.
While it is preferred that the detector is placed at the output plane of the correlator such that it receives a ....
. , -: ~
... . . . .
21~S92~
display of a correlation signal, information as to the correlation of the input and reference fingerprints is available at the output of the filter while the signal remains in the Fourier transform domain. Thus, for example, with reference to figure 1, correlation information could be extracted from the verification device 20 if lens 34 of correlator 38 were dispensed with and detector 36 were positioned more closely to the output of the SLM 26.
A second source of identifying biometric information which can be obtained by the verification device is the cluster shape, the relative lengths and widths of the fingers of the subscriber. These can be measured using either electromechanical transducers or optical scanning devicss. This is particularly u~eful when clu~ters of two or more fingers are used as input to the verification device. The probability of cluster shape values correlating with the fingerprint pattern information are extremely remote, thus adding a higher level of security and confidence to the verification.
Data on cluster width and relative finger lengths is stored along with the fingerprint pattern information upon registration of the subscriber. On verification, this added data i9 correlated with the cluster information collected by the position sensing equipment at the fingerprint input screen. This secondary correlation information is then used by the verification algorithm to determine -~hether the card user is in fact the authorized user.
With reference to figure 1, position transducers 40 are ... . . .
.
~ ~ , .. ~. .
placed on the fingerprint input of the verification device. When an individual places his or her fingers on the input screen, the transduce~s will be displaced according to the lengths and widths of the user's fingers. In the case of optical scanning or S detection equipment, the lengths and widths of the fingers are determined wlthout any mechanical displacements of sensors, but through the interruption or modulation of an optical signal which is projected onto the input finger cluster from an appropriate location.
Other modifications will be apparent to those skilled in the art and, therefore, the invention is defined ln the claims.
. . . .
. : ; .~ . , , ~
,,, ,~
Coherent optical pattern recognition has been studied since its introduction by Anthony Vander Lugt in 1964 (see Vander Lugt, A., Siqnal Detection bY ComPlex SPatial Filterinq IEEE Transactions on Information Theory, vol.IT-10 no. 2 p.139, Apr 1964). This important step wa~ made possible by the development of both a power~ul coherent light source (the laser, in 1961) and off axis holography, by Leith and Upatnieks, also in 1964. From this time to the present, optical techniques for pattern recognition have been known to be extremely powerful because of their speed, high discrimination powers, and noise immunity.
The fundamental phenomenon upon which optical correlators rely is the natural ability of a lens to extract a two dimensional Fourier transform of an input signal. The speed at which the lens will perform the transformation is independent of the input signal or object; a complex input is processed ~ust as fast as a simple one. The computation is done completely in parallel, and occurs literally at the speed of light. This speed represents the most significant advantage that optical processors have over digital devices, and is the reason why so much research has been conducted in this field. Optical devices based on the transformation properties of a lens are commonly referred to as Fourier optical signal processors.
Unfortunately, optical correlation devices suffer from some severe practical difficulties which have prevented them from becoming useful in the real world. In order to exploit the processing speed of the optical technique, input data must be introduced at an extremely high rate, and of course the output must be removed just as fast. Furthermore, in order to compare one pattern with another, a reference or target pattern must be input as well. This reference data is input in the form of a correlation filter, and must be physically located at the transform plane of the correlator. These two input requirements nece~sitate fast, high resolution devices which can compete with the ~nformation bandwidth capabilities of modern digital hardware.
This data transfer problem has always presented a signiflcant bottleneck in Fourier optical signal processing device~. However, this problem is not relevant in applications where high speed, high volume data processing is not a necessity.
Fingerprint verification iq ~ust such an application. In fingerprint verification a device reads recorded information about a reference fingerprint pattern and usQs this information to determine or verify that the reference pattern and an input pattern belong to the same individual. There is no requirement to compare a fingerprint with a large number of reference ~ - .- ~ . , - ... .. ,, .. . ~ , ~105926 patterns, nor is there any need to compare many input patterns with a single reference. A single input pattern must be compared with a single reference pattern, or possibly with a small number of reference patterns. ~revious attempts at optical fingerprint S pattern verification using Fourier techniques have made use of holographic matched filter (HMF) as the correlation filter. The HMF is typically processed on a card which is carried by the user.
However, a disadvantage of optical processing systems which does affect fingerprint verification utilising a Hl~F as the correlation filter is that the position and orientation of the correlation filter is extremely critical to the correct operation of the system. This positional tolerance is on the order of microns, and the rotational orientation must be within one or two d0grees of the optimal position. The mechanical positioning problem i5 therefore extremely difficult, requiring high resolution devices and extremely accurate alignment techniques.
Further, a serious systemic disadvantage is the inflexibility imposed by the holographic matched filter. The necessity of having a holographic matched filter precludes the possibility of transferring filter data by any means other than the physical transport of the original or its copies (which must be fabricated optically as well). The matched filter must be physically present in the frequency plane of the correlator, and there is currently no electronic device which will provide the resolution required to represent a holographic matched filter.
This invention seeks to overcome drawbacks of known ., , , :. . ~ , , : . , , , -... . ~ ,. . ~ . . ; , ,.
fingerprint pattern recognition devices.
SummarY of the Invention According to the present invention, there is provided a fingerprint verification device comprising the following: a source of light for providing an illuminating be~m along a beam path; input means located in the beam path including means to receive at least one fingerprint of an individual (either by the individual placing his finger(s) on the input means or by placing a fingerprint(s) recording in the input means) for modulating the ~eam with data from at least one fingerprint to form an information beam; optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plans; a spatial light modulator fixed in said transform plane; a card reader for reading encoded data from a card representing a correlation filter to be written to said spatial light modulator; means for receiving ~aid encoded data from said card reader and for writing a correlation filter to said spatial light modulator in accordance with said encoded data; verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between the Fourier transform of a reference fingerprint(s) which is derived from filter information contained on a card and written to the spatial light modulator, and the Fourier transform of the fingerprint(s) input.
According to another aspect of the invention, there is provided a method of generating a correlation filter for the . s . . . . -, , :. . . -fingerprint verification device comprising the steps of: forming a correlation filter from each of said at least two fingerprints of an individual which said input means is designed to receive;
superposing each said formed correlation filter to record a correlation filter.
According to another aspect of the invention, there is provided a method of generating a correlation filter for the fingerprint verification device comprising the steps of:
recording a correlation filter from a cluster of all of said at least two fingerprints of an individual which said input means is designed to receive.
Brief Description of the Drawin~s In the drawings which illustrate example embodiments of the invention:
figure 1 i9 a schematic diagram of a verification device made in accordance with this invention; and figure 2a, 2b and 2c are schematic views which illustrate alternate optical corxelation arrangements.
Detailed Description of the Preferred Embodiments In overview, recognising that the filter positioning problem could be solved by permanently fixing and aligning the filter in the proper position at the transform plane, a spatial light modulator ~SLM) is introduced into the transform plane and is fastened there at the time of manufacture. Correlation , 25 filters can then be electronically written onto the SLM, avoiding : . . , , , . ., -the need to mechanically position and align a filter for every correlation operation.
An electronically addressable filter in a correlator allows filter design and calculation to be carried out electronically with a computer. This fact alone makes the verification system far superior to one which utilizes a standard holographic matched filter. Computer generated filters need not be recorded using holographic techniques, which require sophisticated equipment such as lasers and vibration controlled environments. ~olographic matched filters require a holographic--capable recording medium, and finally a stable filter medium which when fixed onto a card to be carried by the user must be able to withstand the rigours of wallets and purses, hot sunlight, freezing temperatures, and spilt coffee and alcohollc beverages. Storlng the filter in a digitally encoded form permits the use of cards carrying encoded data, such as "smart cards", or central databases for filter data storage and retrieval.
Computer generated filters can be far more flexible ln the types of operations which can be performed on an input.
These filter algorithms can be very difficult if not practically impossible to implement optically (holographically). The various filter implementations can optimize the behaviour of the correlator with respect to noise performance, distortion invariance, and discrimination sensitivity. This is accomplished through careful analysis of the recognition problem to be solved.
The filter algorithm can then be tuned to the particular input, , . ~ . ~ ' ' ' . ! . , 2~5926 conditions, and application.
Fingerprint verification is extremely well suited to this application's specific tunability of the computer generated filter. The device has only to compare two sets of fingerprint patterns, and will not be required to differentiate between other types of objects. This allows the filters to be optimized to operate on specific characteristic features of fingerprint patterns which will provide the maximum amount of discrimination.
Distortion and noise characteristics can be studied and accounted for in the filter.
The introduction of an SLM upon which a correlation filter is written opens up a far broader range of useful applications, and makes the system realizable with current technology. The electronic filter data xepresentation permits easy communication of the data from one location to another, and eliminate the severe problems a~sociated with recording, manufacturing, alignment, and reliability of holographic matched filters.
The verification is performed through a hybrid optical/electronic correlation device. The actual correlation is performed optically, while the correlation signal is analyzed electronically. The ability to read stored reference pattern information is made possible through the introduction of a spatial light modulator (SLM3 at the filter plane of the optical correlator.
In practice, with reference to figure 1, a subscriber (an authorized user of a service which utilizes the fingerprint . . . : , . .
. ~ ...: . . . : . :
- . ; . . . .
. . ., . . , :.
2~05926 verification technology) inserts her card 10 into the fingerprint verification device (the verifier) 20, enabling the encoded fingerprint filter information to be txansferred by a reader 22 from the card to the processor 24. This unique information is used by the processor to create an optical filter which is transferred onto the SLM 26, which is contained within the verification device 20. The subscriber then places her fingers on the input screen 28 of the verifier, and the system performs an optical correlation operation, which compares the subscriber's fingerprint(s) with those represented by the encoded fingerprint information stored on the subscriber's card. The output from the SLM is an optical signal whose characteristics depend on the degree of correlation between the input and reference ingerprints. This optical signal, after passing through lens 34 and onto the optical detector 36 where its intensity digtribution i9 detected by the optical detector 36, is analyzed by the processor 24 which decides whether the output signal represent~ a sufficient correlation to indicate a match between the input and reference signals. A positive verification results in the processor 24 allowing the subscriber access to the desired service.
The verification procedure will occur very quickly, on the order of one second or less. The time will be primarily taken up by the various electronic systems which make up the device. The optical correlation operation will occur in several nanoseconds (several billionths of a second), the time required for light to travel from the input to the output plane. The i g verification operation should take no more time than is currently required by systems requiring users to enter PINs (personal identification numbers). This makes the device practical and acceptable as a commercial security system for consumer applications such as automated bank teller (ATM) machines and point of sale credit and debit card machines, while providing tremendously improved security against unauthorized use of such services. The high discrimination capability of Fourier optical pattern recognition techniques will ensure accurate verification.
The encoded representation of the filter information enables the incorporation of a card as an integral component of the security system. Encoded information on a card provides a reliable and readable storage medium for the reference pattern information. The same card can also carry information relevant to the services and privilege~ which the particular card entitles the holder.
The pattern verification is carried out by an optical correlator indicated generally at 38 and which comprises lenses 32 and 34, along with the SLM 26 and a detector 36.
Alternatively, the lenses could be replaced by holographic optical elements which also perform a Fourier transform. The method by which the fingerprint pattern is introduced to the correlator 38 is dependent on the input system, indicated generally at 29. The function of the input system is to generate a high contrast, noise and distortion free image of a live fingerprint(s) pattern and deliver it to the input of the optical correlator 38. The input system 29 may be an all--optical system ...... , . ~, .. . . .. .... .. . . . . . ...... . .. . .
.
.. .. .. . . ..
1 o or a hybrid opto--electronic system which would consist of image capture and display devices. The electronic system has the added capabilities of introducing previously recorded and stored input patterns as well as preprocessing of the input pattern. As illustrated in figure 1, the input system 29 comprises one or more lasers 30 along with an expander lens 31 and a collimator lens 33 to illuminate a prism 35 with a beam 37. One face of the prism forms the input screen 28. In an alternative input system, laser 30 may be replaced with an incoherent light source.
An optical correlator requires some degree of coherence in the optical signal as it passes through the system. Coherence i5 described in terms of spatial and temporal characteristics, and assumes a significant degree of polarization of the light.
Spatial coherence describe~ how light at one point on a plane perpendicular to the optical axis correlates in phase to light at another point on the plane. Temporal coherence relates the correlation of light at one point in space but at different points in time. If a broad band optica~ source is utilized, then ~patial and temporal coherence can be controlled by using source encoding and/or diffraction gratings respectively. Multichannel correlators use a diffraction grating to spatially isolate each wavelength at the filter plane. An alternative to a white light source is a series of lasers, each with a unique wavelength, which provides the advantage of higher spectral power. The SL~
based correlator described here will function with any of the possible light sources which are acceptable for Fourier optical processing devices.
;
~ .
210~926 An all--optical input system will utilize the principle of total internal reflection (~IR) to read the pattern formed by the furrows of the input fingerprint pattern. A furrow will create an air space over the surface of a glass screen, allowing 5light which is internally reflecting from the interior surface of the screen to proceed unimpeded. Ridges, however, will be in contact with the surface, where they will scatter and absorb a portion of the illuminating light. This effect is known as frustrated total internal reflection (FTIR). When the TIR light 10signal is observed, the primary signal will be due to the furrow pattern.
If the scattered light from the ridge pattern is observed, there may be significant advantages in terms of noise immunity from latent fingerprints. The scattered light signal 15will be weaker than the TIR signal from the furrows. Despite thls cost in illumlnation efficiency, the improved signal to noise ratio (SNR) of the FTIR signal from the ridges will most likely result in improved performance of the verification system in a real world environment.
20For both of the preceding all--optical input schemes, the light signal from the input pattern enters the optical correlator directly, without further interference or signal conditioning from any active electronic devices.
The alternative to the all--optical input system is an 25electronlc imaging and display system which captures the image of the input pattern and displays it on a ~evice such as an SLM.
~: The image capture may be performed by an optical array detector, : ~ :
210~926 scanner or similar device. The image can then be displayed at the input of the optical correlator 38 by illuminating the display device, either in transmission or reflection. This scheme permits preprocessing of the input pattern. PreproceSsing can provide added flexibility and power to an optical processor.
Alternatively, an optically addressable SLM can be used at the correlator input. This type of device allows a spatial pattern of light on one side of the SLM to affect the reflectivity on the other. Optically addressable SLMs are commonly used in optical signal processing devices to perform incoherent to coherent light conversion.
Input pattern illumination for the electronic imaging and display input may be the same as for the all--optical scheme;
TIR or FTIR could be exploited to create a high contrast image of the pattern.
The optical signal carrying the fingerprint pattern information proceeds into the correlator 38 and is transformed by the Fourier transform lens 32. Such a lens will perform an optical transform of the input pattern and display the transform at the Fourier transform plane of the correlator. The Fourier transform plane is typically located at the focal plane of the transform lens 32. The SLM 26 is positioned at the transform plane, and the optical transform of the input pattern passes through the SLM. A correlation filter is written onto the SLM, modulating the optical transform through a multiplicative operation with the function represented on the SLM. The second Fourier transform lens 34 i5 positioned in such a way that the . -210S92~
output of the SLM is re--transformed, resulting in the display of a correlation signal at the output plane of the correlator.
The correlation signal represents the transform of the product between the correlation filter written to the spatial light modulator - i.e., the reference fingerprint represented by the data read from the card 10 - and the Fourier transform of the fingerprints of the subscriber at the input screen 28. The correlation signal is usually accompanied by DC, convolution, and noise signals. These additional signals and their relative strengths depend on the correlator architecture and the filter characteristics.
The actual function represented on the SLM for correlation with a reference pattern can be arrived at through numerous method~, and they are described by various names lndicating their particular attributes. This flexibility is in fact a tremendous advantage of electronic filter calculation.
Whatever the algorithm, all the filters are designed to produce a corr~lation signal at the output of the correlator.
An array detector 36 at the output plane of the correlator intercepts the correlation signal, converting it to its electronic analog. This electronic signal i~ then captured, stored and analyzed by the processor 24 (electronic computer).
Through analysis of specific characteristics of the correlation ; signal the processor decides whether or not there is sufficient correlation between the input and reference patterns to warrant - a positive verification. The decision of the processor may be signalled to a host system, which then takes the appropriate : . - . -:. .
. . ~ . - - ,: -action.
The SLM is a device which can spatially m~dulate (in two dimensions) the amplitude and/or phase of light in either transmission or reflection. The spatial resolution of an SLM
varies widely and depends on the type of SLM. SLMs can be divided into two main categories, optically addressable and electrically addressable. In the optically addressable SLM
(OASLM) a pattern of light (coherence is irrelevant) is written onto the device, imposing an associated spatial modulation of a second (coherent) beam of light. In the electrically addressable SLM (EASLM) the magnitude of modulation of individual cells or pixels are controlled electrically, with each pixel having its own unique address.
O particular intere~t is the liquid crystal television or dlsplay (LCTV), These devices are widely used as flat panel displays for portable computers and calculators, control panel displays on equipment and instrumentation, and as video and pro~ection display devices for small televisions and projectors.
LCTVs are relatively inexpensive, have satisfactory spatial resolution, and can be made to modulate light in amplitude and/or phase. LCT~s have proven themselves to be mas6 producible and reliable, and have minimal power requirements.
Current technology does not provide the optical system designer with an SLM possessing the spatial resolution capabilities of a holographic recording emulsion, but the resolution is sufficient for correlation operations in an optical data processing device. However, the low spatial resolution - - .
.,, . , ., . . , . .. . ... .- .. - .. ~
210~926 results in a low diffraction efficiency, while the information content required to implement a classical matched filter, even with a 128 x 128 pixel S1M, i8 not small. Advances in filter algorithms has resulted in superior approaches to filter implementations which do not simply attempt to implement a low resolution holographic matched filter.
It must be noted that a holographic matched filter employs a carrier frequency (created by the reference beam) to encode the complex amplitude information on an amplitude only format. This encoding scheme stores both amplitude and phase information, which naturally requires a high information storage capacity from the medium, i.e. high resolution. If the filter modulates only the phase of the light, not only is the efficiency improved (there are no absorption losses in this case), but the information content ls sub3tantially reduced. Such a filter is called a phase only filter (POF), and hag been shown to perform a~ well or even better than a matched filter.
Even further ~aving~ in data storage can be had if the filter function i~ binarlzed. This proce~s conver~s a continuous pha~e function lnto a two level function. There are once again numerous algorithms for binarization, but the data storage requirements can be seen to be extremely minimal. Such a filter is called a binary phase only filter (BPOF), and has also been shown to perform extremely well in spite of the reduced information content.
A 3ignificant advantage of the SLM as a filter is the flexibility afforded by its ability to rapidly change filters ,,, ", . ".,~ . "' 1 " ,, , ,, ,, . ~,, ., . . , :, " :, ", ' ,' ., " , . . .
. . . :. :, ,. ~
- ~lD~926 without mechanically replacing the filter. An SLN can therefore be incorporated into the correlator and aligned at the time of manufacture, avoiding the difficulty of mechanically aligning a HMF in the transform plane at every verification operation.
Holographic filter alignment is known to be very critical for correct operation of a HMF correlator, requiring precision on the order of microns. This places a tremendously high emphasis on the mechanical accuracy and reliability of a positioning system.
The operation of the SLM can be in amplitude, phase, or full complex (amplitude and phase) modulation. The modulation can also be in binary form, as opposed to continuous modulation.
Binary modulation, as previously stated, has the advantage of greatly reduced data storage requirements.
~nother advantage of the SLM--based correlator is that the design and fabrication of the filter can be performed electronically, introducing far more flexibility in filter design. A computer generated filter (CGF) can implement operations which can be very difficult or practically impossible to perform with an optically recorded filter. This make~ filter design and testing much faster and more effective. For example, the SLM can compen~ate for any misalignment in the fingerprint on the input screen of the verifier. This may be accomplished by electronically rotating the ~ilter on the fixed SLM. The actual degree of misalignment does not need to be known since the filter on the SLM may be electronicàlly rotated until the maximum output signal is found. Alternatively, the filter on the SLM
could be made rotationally invariant. Furthermore, the pattern :., , .. , , ., ........... . . - .. .. :.... . ......... .
,, ; . , . . . . - ~ , .
- 17 _ on the filter may be electronically made larger or smaller to accommodate scale changes of the input fingerprint. Modern digital techniques permit preprocessing of the reference fingerprint pattern to optimize the correlators performance in terms of expected noise and input distortions. Significant advances in filter algorithms provide the system designer with a large number of starting points from which to start developing an application specific filter.
Electronic or digital storage of the reference fingerprint filter information avoids the practical difficulties associated with holographic recordings and recording media which must be dealt with in a HMF--based sy~tem. Instead of the holographic recording medium on the card, the information required to write the filter onto the SLM is stored on a card such as a smart card. The verifier reads this information from an inserted card and writes the pre~cribed correlation filter onto the SLM.
Encoded fingerprint filter information storage represent~ an enormous advance over the HMF in the collection, processing and distribution of fingerprint data. The enrolment procedure can now be an easily di~tributed operation.
Fingerprint scanners will simply digitize subscribers' fingerprints and relay the information to a filter-generating computer. The ease with which digital information can be transmitted makes the actual location of the computer irrelevant.
Moreover, advances in computer and communications technologies make the po~sibilities of a near real time enrolment a realistic 21~S926 possibility. With such a system a new subscriber would insert his newly issued blank card into the enrolment device, apply his fingerprints to the screen, and remove the initialized card ~everal seconds later, thus completing the enrolment. The use of two or more fingerprints (a cluster) offers significant advantages over the use of just a single digit. Two approaches are possible. The first involves the digitization of the cluster as a single image, from which a filter is generated just as it would be generated for a single digit. (Digitization and filter generation are independent of the image content.) There is obviously more identifying information in a cluster than in a single digit, resulting in a filter which has the same amount of information ~the same number of data bits) but with more unlque information content. This information now includes relative finger position (i.e. relative length and orientation) because of the ~patial relationships between the features of one digit of the cluster with the all the features on the other digit(g) .
A second approach to the use of a cluster is to digitize each finger independently. Independent filters could be generated for each finger in the cluster, or a single composite filter could be derived. The case of two or more independent filters clearly represents more information than a ~ingle filter. At verification, the cluster is applied to the input of the verification device which in this application uses one or more sources of light. The card, holding independent correlation filter data for each digit in the cluster, is , ,' ' ~ .. ~ " ~' , ''.,' ,', ':' ' ' :: . . , : . . . .
-: . : . -210592~
inserted into the card reader as usual. The verification device then displays each filter sequentially, verifying the presence of each digit in the cluster. A unique feature of this approach is that the correlation peak location at the output is dependent on the actual location of the target digit on the input screen.
By tracking the correlation peak location for each filter, the cluster shape (i.e. the relative finger lengths and widths) can be reconstructed. This provides additional biometric information to the system which can be used to improve system performance beyond what would be possible with only a single digit.
Alternatively, a single composite filter can be constructed from the digitized images of each finger in the cluster. This will provide a single filter capable of correlating with each digit in the cluster, and also provide relative digit location in the lnput plane of the verification device.
Encoded filter data also provides a convenient format for information storage and retrieval. The verification technology then becomes applicable where ever a centrally accessible database of authorized subscribers is desirable. This type of security system would not require the ~ubscriber to carry a filter card at all. For example, the verifier could access the subscriber's encoded fingerprint information from a central data~ase by cross referencing with the user's P.I.N. entered on a keyboard. The fingerprint of the user is then correlated using~
a filter written onto the SLM with the centrally stored pattern information.
If the database is small enough and the S~M fast . :. : . : . ..... .. ~ . , :
, ' , .. !.. : . . ' . , , ~ . . . . .
: ' ,, .. . '' ~ . ' ' , ' '. ', , ' ' ,' ': ' '' . '' , '' ' . ' .' . ~ ,.
enough, a match for an input fingerprint can be searched for by sequentially scanning through the database. This approach may be desirable for access to high security installations or systems, etc., where there are relatively few authorized per30nnel. Current LCTV devices have scan rates of 60 Hz., allowing a search through a thirty record database in half a second. The ~peed can be expected to improve with advances in technology.
A further advantage of the digital filter format is the flexibility it offers for upgrading and improving the system.
Improved filter designs are readily implemented by the filter generating computer. Existing cards can be upgraded at the subscriber's convenience at any enrolment facility. Upgrading of S~M drivers and electronics in the verifiers can be tolerated beaause o~ the electronic nature of these devlces. Advances in relevant technology can be exploited, yet backwards compatibility can be always be ensured through software.
It should be noted that the optical arrangement for the correlator 38 illustrated in figure 1 is by way of example only.
Other optical arrangements are known. In this regard, reference is made to figures 2a, 2b, and 2c which shows a few of the known alternate arrangements for an optical correlator with P1 indicating the input, P2 the correlation filter, and P3 the detector. It will be recognised that figure 2b illustrate~ the optical arrangement for the correlator illustrated in figure 1.
While it is preferred that the detector is placed at the output plane of the correlator such that it receives a ....
. , -: ~
... . . . .
21~S92~
display of a correlation signal, information as to the correlation of the input and reference fingerprints is available at the output of the filter while the signal remains in the Fourier transform domain. Thus, for example, with reference to figure 1, correlation information could be extracted from the verification device 20 if lens 34 of correlator 38 were dispensed with and detector 36 were positioned more closely to the output of the SLM 26.
A second source of identifying biometric information which can be obtained by the verification device is the cluster shape, the relative lengths and widths of the fingers of the subscriber. These can be measured using either electromechanical transducers or optical scanning devicss. This is particularly u~eful when clu~ters of two or more fingers are used as input to the verification device. The probability of cluster shape values correlating with the fingerprint pattern information are extremely remote, thus adding a higher level of security and confidence to the verification.
Data on cluster width and relative finger lengths is stored along with the fingerprint pattern information upon registration of the subscriber. On verification, this added data i9 correlated with the cluster information collected by the position sensing equipment at the fingerprint input screen. This secondary correlation information is then used by the verification algorithm to determine -~hether the card user is in fact the authorized user.
With reference to figure 1, position transducers 40 are ... . . .
.
~ ~ , .. ~. .
placed on the fingerprint input of the verification device. When an individual places his or her fingers on the input screen, the transduce~s will be displaced according to the lengths and widths of the user's fingers. In the case of optical scanning or S detection equipment, the lengths and widths of the fingers are determined wlthout any mechanical displacements of sensors, but through the interruption or modulation of an optical signal which is projected onto the input finger cluster from an appropriate location.
Other modifications will be apparent to those skilled in the art and, therefore, the invention is defined ln the claims.
. . . .
. : ; .~ . , , ~
,,, ,~
Claims (14)
1. A fingerprint verification device comprising the following:
- a source of light for providing an illuminating beam along a beam path;
- input means located in the beam path including means to receive at least one fingerprint of an individual for modulating the beam with data from the at least one fingerprint to form an information beam;
- optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plane;
- a spatial light modulator fixed in said transform plane;
- a card reader for reading encoded data from a card representing a correlation filter to be written to said spatial light modulator;
- means for receiving said encoded data from said card reader and for writing a correlation filter to said spatial light modulator in accordance with said encoded data;
- verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between a correlation filter written to said spatial light modulator and the Fourier transform of at least one fingerprint of an individual received at said input means.
- a source of light for providing an illuminating beam along a beam path;
- input means located in the beam path including means to receive at least one fingerprint of an individual for modulating the beam with data from the at least one fingerprint to form an information beam;
- optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plane;
- a spatial light modulator fixed in said transform plane;
- a card reader for reading encoded data from a card representing a correlation filter to be written to said spatial light modulator;
- means for receiving said encoded data from said card reader and for writing a correlation filter to said spatial light modulator in accordance with said encoded data;
- verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between a correlation filter written to said spatial light modulator and the Fourier transform of at least one fingerprint of an individual received at said input means.
2. The fingerprint verification device of claim 1 wherein said input means comprises an input screen for the reception of at least one finger of an individual.
3. The fingerprint verification device of claim 1 wherein said input means comprises means to receive a recording of at least one fingerprint of an individual.
4. The fingerprint verification device of claim 1 wherein said optical Fourier transform means comprises a lens.
5. The fingerprint verification device of claim 1 wherein said optical Fourier transform means comprises a holographic optical element.
6. The fingerprint verification device of claim 4 or claim 5 wherein said verification means is positioned so that any Fourier transform signal at said spatial light modulator is re-transformed at said verification means.
7. The fingerprint verification device of claim 6 wherein said verification means comprises a detector in said beam path and a processor responsive to the output of said detector.
8. The fingerprint verification device of claim 7 wherein said means for writing a correlation filter to said spatial light modulator includes a processor responsive to the output of said card reader and operatively connected to the input of said spatial light modulator.
9. A fingerprint verification device comprising the following:
- a source of light for providing an illuminating beam along a beam path;
- input means located in the beam path including means to receive a recording of at least one fingerprint of an individual for modulating the beam with data from the recording of at least one fingerprint to form an information beam;
- optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plane;
- a spatial light modulator fixed in said transform plane;
- a card reader for reading encoded data from a card representing a correlation filter to be written to said spatial light modulator;
- means responsive to said card reader for receiving said encoded data and for writing a correlation filter to said spatial light modulator in accordance with-said encoded data;
- verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between a correlation filter written to said spatial light modulator and the Fourier transform of the recording of at least one fingerprint of an individual received at said input means.
- a source of light for providing an illuminating beam along a beam path;
- input means located in the beam path including means to receive a recording of at least one fingerprint of an individual for modulating the beam with data from the recording of at least one fingerprint to form an information beam;
- optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plane;
- a spatial light modulator fixed in said transform plane;
- a card reader for reading encoded data from a card representing a correlation filter to be written to said spatial light modulator;
- means responsive to said card reader for receiving said encoded data and for writing a correlation filter to said spatial light modulator in accordance with-said encoded data;
- verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between a correlation filter written to said spatial light modulator and the Fourier transform of the recording of at least one fingerprint of an individual received at said input means.
10. A fingerprint verification system comprising the following:
- a source of light for providing an illuminating beam along a beam path;
- input means located in the beam path including means to receive at least one fingerprint of an individual for modulating the beam with data from the at least one fingerprint to form an information beam;
- optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plane;
- a spatial light modulator fixed in said transform plane;
- a card storing a correlation filter in encoded data form;
- a card reader for reading encoded data from said card;
- means for receiving said encoded data from said card reader and for writing a correlation filter to said spatial light modulator in accordance with said encoded data;
- verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between a correlation filter written to said spatial light modulator and the Fourier transform of at least one fingerprint of an individual received at said input means.
- a source of light for providing an illuminating beam along a beam path;
- input means located in the beam path including means to receive at least one fingerprint of an individual for modulating the beam with data from the at least one fingerprint to form an information beam;
- optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plane;
- a spatial light modulator fixed in said transform plane;
- a card storing a correlation filter in encoded data form;
- a card reader for reading encoded data from said card;
- means for receiving said encoded data from said card reader and for writing a correlation filter to said spatial light modulator in accordance with said encoded data;
- verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between a correlation filter written to said spatial light modulator and the Fourier transform of at least one fingerprint of an individual received at said input means.
11. The fingerprint verification device of claim 10 wherein said input means is for receiving at least two fingerprints of an individual and wherein said encoded data comprises a correlation filter which is the superposition of a separate correlation filter for each of said at least two fingerprints.
12. The fingerprint verification device of claim 10 wherein said input means is for receiving at least two fingerprints of an individual and wherein said encoded data comprises a correlation filter which is a single image of said at least two fingerprints.
13. A method of generating a correlation filter for a fingerprint verification device of the type having:
- a source of light for providing an illuminating beam along a beam path;
- input means located in the beam path including means to receive at least two fingerprints of an individual for modulating the beam with data from the at least two fingerprints to form an information beam;
- optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plane;
- a spatial light modulator fixed in said transform plane;
- a card reader for reading encoded data from a card representing a correlation filter to be written to said spatial light modulator;
- means for receiving said encoded data from said card reader and for writing a correlation filter to said spatial light modulator in accordance with said encoded data;
- verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between a correlation filter written to said spatial light modulator and the Fourier transform of at least two fingerprints of an individual received at said input means, comprising the steps of:
- forming a correlation filter from each of said at least two fingerprints of an individual which said input means is designed to receive;
- superposing each said formed correlation filter to record a correlation filter.
- a source of light for providing an illuminating beam along a beam path;
- input means located in the beam path including means to receive at least two fingerprints of an individual for modulating the beam with data from the at least two fingerprints to form an information beam;
- optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plane;
- a spatial light modulator fixed in said transform plane;
- a card reader for reading encoded data from a card representing a correlation filter to be written to said spatial light modulator;
- means for receiving said encoded data from said card reader and for writing a correlation filter to said spatial light modulator in accordance with said encoded data;
- verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between a correlation filter written to said spatial light modulator and the Fourier transform of at least two fingerprints of an individual received at said input means, comprising the steps of:
- forming a correlation filter from each of said at least two fingerprints of an individual which said input means is designed to receive;
- superposing each said formed correlation filter to record a correlation filter.
14. A method of generating a correlation filter for a fingerprint verification device of the type having:
- a source of light for providing an illuminating beam along a beam path;
- input means located in the beam path including means to receive at least two fingerprints of an individual for modulating the beam with data from the at least two fingerprints to form an information beam;
- optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plane;
- a spatial light modulator fixed in said transform plane;
- a card reader for reading encoded data from a card representing a correlation filter to be written to said spatial light modulator;
- means for receiving said encoded data from said card reader and for writing a correlation filter to said spatial light modulator in accordance with said encoded data;
- verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between a correlation filter written to said spatial light modulator and the Fourier transform of at least two fingerprints of an individual received at said input means, comprising the steps of:
- recording a correlation filter from a cluster of all of said at least two fingerprints of an individual which said input means is designed to receive.
- a source of light for providing an illuminating beam along a beam path;
- input means located in the beam path including means to receive at least two fingerprints of an individual for modulating the beam with data from the at least two fingerprints to form an information beam;
- optical Fourier transform means in the beam path for providing a Fourier transform of the information beam in a transform plane;
- a spatial light modulator fixed in said transform plane;
- a card reader for reading encoded data from a card representing a correlation filter to be written to said spatial light modulator;
- means for receiving said encoded data from said card reader and for writing a correlation filter to said spatial light modulator in accordance with said encoded data;
- verification means in the beam path at the output of said spatial light modulator for sensing a correlation signal which is the transform of the product between a correlation filter written to said spatial light modulator and the Fourier transform of at least two fingerprints of an individual received at said input means, comprising the steps of:
- recording a correlation filter from a cluster of all of said at least two fingerprints of an individual which said input means is designed to receive.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US94348492A | 1992-09-11 | 1992-09-11 | |
| US07/943,484 | 1992-09-11 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA2105926A1 true CA2105926A1 (en) | 1994-03-12 |
Family
ID=25479744
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA002105926A Abandoned CA2105926A1 (en) | 1992-09-11 | 1993-09-10 | Fingerprint verification system |
Country Status (2)
| Country | Link |
|---|---|
| CA (1) | CA2105926A1 (en) |
| GB (1) | GB2270586B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5481621A (en) * | 1992-05-28 | 1996-01-02 | Matsushita Electric Industrial Co., Ltd. | Device and method for recognizing an image based on a feature indicating a relative positional relationship between patterns |
| JPH0981725A (en) * | 1995-09-08 | 1997-03-28 | Hamamatsu Photonics Kk | Human collating device |
| JPH09147115A (en) * | 1995-11-20 | 1997-06-06 | Hamamatsu Photonics Kk | Personal collating device |
| GB2309329A (en) * | 1996-01-18 | 1997-07-23 | John Gerald Bullard | Fingerprint recognition |
| US6463416B1 (en) | 1996-07-15 | 2002-10-08 | Intelli-Check, Inc. | Authentication system for identification documents |
| GB9709883D0 (en) * | 1997-05-16 | 1997-07-09 | Ncr Int Inc | User verification system |
| GB2332545B (en) * | 1997-12-17 | 2002-03-13 | Marconi Gec Ltd | Radar return signal signature analysis |
| US5995014A (en) * | 1997-12-30 | 1999-11-30 | Accu-Time Systems, Inc. | Biometric interface device for upgrading existing access control units |
| US7708189B1 (en) | 2002-05-17 | 2010-05-04 | Cipriano Joseph J | Identification verification system and method |
| US7860318B2 (en) | 2004-11-09 | 2010-12-28 | Intelli-Check, Inc | System and method for comparing documents |
| US10373409B2 (en) | 2014-10-31 | 2019-08-06 | Intellicheck, Inc. | Identification scan in compliance with jurisdictional or other rules |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB8609700D0 (en) * | 1986-04-21 | 1986-05-29 | Gen Electric Co Plc | Optical database correlator |
| US5050220A (en) * | 1990-07-24 | 1991-09-17 | The United States Of America As Represented By The Secretary Of The Navy | Optical fingerprint correlator |
| JP3378032B2 (en) * | 1992-08-28 | 2003-02-17 | 浜松ホトニクス株式会社 | Person verification device |
-
1993
- 1993-09-10 CA CA002105926A patent/CA2105926A1/en not_active Abandoned
- 1993-09-13 GB GB9318889A patent/GB2270586B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| GB2270586A (en) | 1994-03-16 |
| GB2270586B (en) | 1995-11-08 |
| GB9318889D0 (en) | 1993-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR100874294B1 (en) | Authentication system and method | |
| US8150033B2 (en) | Optical security system using fourier plane encoding | |
| US5712912A (en) | Method and apparatus for securely handling a personal identification number or cryptographic key using biometric techniques | |
| US5920058A (en) | Holographic labeling and reading machine for authentication and security applications | |
| US4876725A (en) | Method and apparatus for fingerprint verification | |
| US5003600A (en) | Diffraction gratings used as identifying markers | |
| US6111671A (en) | Match filters for a real time fingerprint sensor and verification system | |
| WO1997005594A1 (en) | Holographic method for encrypting and decrypting information using a fingerprint | |
| WO1996041297A1 (en) | Hybrid optical-digital method and apparatus for fingerprint verification | |
| CA2105926A1 (en) | Fingerprint verification system | |
| Millan | Advanced optical correlation and digital methods for pattern matching—50th anniversary of Vander Lugt matched filter | |
| Alam et al. | Real-time fingerprint identification | |
| Brasher et al. | Incoherent optical correlators and phase encoding of identification codes for access control or authentication | |
| KR100402397B1 (en) | method for manufacturing the security card using a hologram and system for authentication of the security card | |
| Javidi | Optical spatial filtering for image encryption and security systems | |
| KR20030019988A (en) | Device for recording personal id information on a hologram, hologram card having the hologram printed thereon and device for reading the hologram card | |
| Markov et al. | Random encoded reference beam for secure data storage in a holographic memory | |
| Markov et al. | Volume hologram with random encoded reference beam for secure data encryption | |
| KR20040057210A (en) | Apparatus and Method for Recognizing Hologram Fingerprint | |
| Horner et al. | Correlation System for Security Validation and Verification Using An Encoded Phase Mask | |
| Javidi | Optical processing and pattern recognition applied to security and anticounterfeiting | |
| Javidi et al. | V: Pattern Recognition with Nonlinear Techniques in the Fourier Domain | |
| Alam et al. | Joint transform correlation for fingerprint identification | |
| Cederquist et al. | Analytic design of optimum holographic optical elements | |
| Muravsky et al. | Comparative analysis of optical and hybrid joint tranform correlators for security applications |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| FZDE | Discontinued |