 |
Description  |
|
|
BACKGROUND
1. Field of the Invention
This invention relates to methods and apparatus for identifying human
beings, specifically to an improved method for reading and analyzing
fingerprints. This invention is also useful for acquiring better quality
fingerprints without the need of ink or lubricating gel for law
enforcement applications.
2. Description of Prior Art
Fingerprints have been a safe means for uniquely identifying human beings
for over one hundred years. So far there has not been a single instance of
two human beings having the same fingerprints. Fingerprints have been used
for human identification by law enforcement for a long time. Recently, as
technology has become available, fingerprints have been used for real-time
human being identification in access control applications, personal
identification (ID), ID verification, in documenting the time of
attendance, identification of intruders, and similar applications.
Despite the long history of fingerprints and the reliable identification
that they offer, it has been very hard to achieve the same level of
identification accuracy with an electronic fingerprint recognition system
as is possible by an experienced fingerprint forensic analyst. The
discrepancy in the level of identification accuracy with the existing
electronic systems is attributed to various reasons. The main reasons
primarily relate to:
(a) the weakness of existing electronic systems to extract the
characteristic points (Minutiae) from a fingerprint reliably and
consistently;
(b) the weakness of the existing electronic scanners to compensate for
natural variations and wear of the fingerprint due to dryness from dry and
cold weather, the cracking due to hard water, soap, detergent, and contact
with abrasive surfaces;
(c) the weakness of existing software to compensate for the variations
caused by the conditions enumerated in (b) above.
Substances that have been used traditionally by law enforcement to
compensate for the tear and wear of the fingerprint of a human being
include ink and clear gel. Both means of compensation are messy, leave
residues on the finger, and thus completely inadequate for real-time
fingerprint capturing for repetitive use of the fingerprint as required in
access control applications, personal ID verification, time of attendance
determination, and for intrusion control, etc.
Traditionally fingerprint identification systems rely on Minutiae to
identify the person to whom the fingerprint belongs to. Minutiae are the
points of termination of a ridge, or the points of bifurcation of a ridge
in a fingerprint. Practically all methods of fingerprint matching from the
Prior Art are based on Minutiae matching. Although Minutiae constitute an
irrefutable means of fingerprint matching, the accuracy of Minutiae
matching is questionable under a wide variety of finger conditions. For
example Minutiae become very unreliable as a means of matching, when a
fingerprint is chapped, dry and, in general, worn out from natural wear.
Chapped, dry, or worn out fingerprints produce random sets of Minutiae
either because ridges become fragmented or because they become very faint
to detect. In either case, Minutiae become an unreliable means of
accurately identifying fingerprints, and thus the identity of a human
being from them.
In our invention, as it will become clear in the next section, the matching
method that we use relies on fusing the information contained in the
Minutiae with information extracted from another orthogonal set of
features that relates to the ridge frequency. By ridge frequency, we imply
the rate of succession between ridges and valleys in the fingerprint image
at any given direction on the image plane. The ridge frequency information
provides information that can be used to discriminate one fingerprint from
another. This information is referred to as Meniscus or Global in the
invention.
By fusion we mean the process of integrating two different sets of data in
a meaningful way. In order to maximize the beneficial results from fusion,
the data sets must be orthogonal in the sense that the information in one
data set compliments the information from the other. This complimentarity
of information is crucial in achieving enhanced performance after fusion
of different data sets.
One important attribute of the Meniscus or Global feature set is that it is
orthogonal to the Minutiae features in the sense that the microscopic
Minutiae structures are not affected by the macroscopic Meniscus/Global
ridge structures. This orthogonality is important in our invention,
because it allows to fuse the information from the two feature sets in a
complimentary fashion, thus enhancing the fingerprint matching accuracy
beyond what is feasible by each method individually. Moreover, the two
feature sets are structurally complimentary, so when one set becomes
unreliable due to finger degradation, the other still maintains a level of
quality that allows to do accurate identification. So, the orthogonality
allows to use both sets synergistically by fusing them, and enhance the
identification accuracy after fusion.
Fusion has been used in other fields and the inventors have many
contributions in the field of fusion. However, fusion has never been
applied before for human being identification by combining information
contained in different orthogonal feature sets from the same fingerprint
with the purpose to reduce or, even eliminate all together, the false
rejection and false acceptance rates when matching one fingerprint against
another. False rejection rate is defined as the rate of failure of a
properly authorized human being. False acceptance rate is defined as the
rate of acceptance of an unauthorized human being. Although fusion of
Minutiae with Meniscus or Global features provides an identification
method far superior in terms of accuracy than any method in the Prior Art,
there are circumstances that a finger may be either extremely chapped,
dry, or worn out from weather, temperature, manual labor, or dirt, that
does not leave a trace (fingerprint) on the scanner to capture. Under
these conditions, no identification method, no matter how sophisticated
can be used for identification: the fingerprint it's not there. It is
precisely this situation that our invention provides the only possible
remedy for: a conditioning apparatus that when in contact with the finger
releases a controlled amount of lubricant, or lubricant mixture, that
rebuilds the fingerprint structure so that it becomes visible to the
scanner. Furthermore, the same conditioning apparatus can be used to
enhance the fingerprint quality so that tighter security levels can be
used in a system that identifies human beings from fingerprints.
OBJECTS AND ADVANTAGES
This patent describes the use of an oil impregnated material or, similarly
an apparatus, that releases a controlled amount of lubricant, or a
lubricant solution with a disinfectant and fragrant, when in touch with an
fingerprint and/or under pressure for conditioning a finger prior to, or
during, or after the process of capturing the fingerprint either
electronically or mechanically on paper or other solid surface. The term
conditioning throughout this invention indicates a process of lubricating
and/or cleaning and/or controlling excessive moisture from a finger. This
invention also describes a new use of an oil impregnated material, the
MICROPOLY, or any variant of it manufactured by PhyMet Inc. of Ohio, or
any similar material or version of the above, that can release a
controlled amount of lubricant, or a mixture of lubricant with fragrant
and/or disinfectant, on the finger, when the material is touched or
subjected to pressure, and can be used to condition the finger prior to
scanning in order to compensate for the wear of the fingerprint caused by
any of the reasons mentioned above. The use of this material in the
real-time fingerprint recognition system for which this patent is filed,
allows to use the fusion method for which this patent is filed to achieve
negligible probability of false acceptance while maintaining very low
false rejection rate. The combination of these two attributes, negligible
false acceptance probability and very low rejection rate, make the system
for which a patent is filed unique and provide the operational assurance
required by the industry to make the use of fingerprints mainstream in
access control, security applications, ID verification, time attendance
application, or any related application where identification of a human
being is desired.
Accordingly, several objects and advantages of our invention are:
(a) to provide a non-intrusive, non-objectionable means for fingerprint
conditioning, whereas the prior art in this area provides non-friendly,
intrusive, and messy substances for finger conditioning;
(b) to provide a fingerprint processing and matching method that, in
conjunction with object (a), guarantees negligible false acceptance
probability at very low false rejection rate, whereas the prior art in
this area has failed to deliver systems that reliably guarantee low false
acceptance and low false rejection rates;
(c) to provide an apparatus that incorporates both objects (a) and (b) to
realize a high integrity, high reliability real-time fingerprint
recognition system, whereas the prior art in this area does not have a
system that provides acceptable means for conditioning fingerprints and
matching techniques to guarantee the low false acceptance and false
rejection rates acceptable in real-time fingerprint recognition systems;
and
(d) to provide an apparatus that can integrate fingerprint scanning
capabilities with keypad technology and card reader technology and the
software to go along with it, so as a security network can easily be built
by daisy-chaining those apparatuses to allow for cost-effective security
solution to realize an onion-like security concept, whereby the outermost
vulnerable perimeter layer is protected by the use of fingerprint
verifiable Personal Identification Numbers (PINs) or card passwords, and
the less vulnerable interior layers may be protected by conventional
keypad or card reader security devices that do not require user ID
verification. The prior art in this area does not offer systems that are
configurable for multiple ID elements, such as PIN, password, and
fingerprints in any desirable combination all operating off the same
database and capable of being connected on the same network.
Further objects and advantages of our invention will become apparent from a
consideration of the drawings and ensuing description.
SUMMARY OF THE INVENTION
In summary, the method and apparatus for accurate measurement of the
identity of human beings is:
1. A method for accurately determining identity of a human being comprising
the steps of:
a. conditioning a finger (100) of said human being by bringing said finger
in contact with a material or device (102) that releases a controlled
amount of lubricant on the surface of said finger containing unique
identifying characteristics;
b. capturing a fingerprint image (110) of said lubricated finger (103)
electronically with a scanner (104) or by leaving an imprint of said
finger mechanically on a surface; and identifying said human being either
by processing said fingerprint image electronically with a computer (106)
or manually by visual inspection.
2. A method for achieving superior identification of a human being
comprising the steps of:
a. extracting Minutiae (172) from said fingerprint image and feature
matching (178) said Minutiae (172) by ranking against other Minutiae
templates in a database;
b. extracting Meniscus or Global ridge frequency features (174) from said
fingerprint image (170) and feature matching (178) by ranking said
Meniscus or Global ridge frequency features (174) by ranking said Meniscus
or Global ridge frequency features against candidate templates from said
finger (103) in said database;
c. fusion (180) of rank scores from said Minutiae (172) features with said
Meniscus or Global (174) ridge frequency feature sets to achieve superior
identification accuracy of said human being.
3. A method for improving the quality of a fingerprint image (126) from a
dry, chapped, or dirty finger (123) of a human being, comprising:
a. use of an oil impregnated material (102), or a similar material
impregnated with lubricant or mixture of lubricant with fragrant and
disinfectant capable of releasing a controlled amount of lubricant on said
finger (123) under contact or pressure;
b. use of a scanner (104) for capturing a high quality unfragmented
fingerprint image (126) possessing unbroken ridge structure of said
conditioned dry, chapped, or dirty finger (125) for identifying said human
being.
4. An apparatus for improving quality of a fingerprint image (126) of a
human being comprising:
a. the use of oil impregnated MICROPOLY material (132), or equivalent
material impregnated with lubricant or mixture of lubricant with fragrant
and disinfectant, in the form of inkless stamp pad (158), pellet (132),
screw-on pellet (162), or an electromechanical dispenser (164), that
includes said pellet (160) of oil impregnated MICROPOLY material (132) and
a plate (168) and spring (116) mechanism to exert pressure on said pellet
(160), each capable of releasing a controlled amount of lubricant on
finger 123) under contact or pressure; and
b. means for capturing a high quality unfragmented fingerprint image (126)
of said conditioned finger (125) for identifying said human being.
5. An apparatus (128) for identification of a human being from fingerprints
comprising:
a. a pellet (132) of oil impregnated MICROPOLY material (132), or
equivalent conditioning material or device to lubricate, clean, deodorize,
or disinfect a finger prior to placing said finger on a scanner (104);
b. a scanner (104) for capturing a fingerprint image (126);
c. means for identifying said human being by extracting features from said
fingerprint image and comparing and matching said features with features
from other fingerprint images in a database.
6. A method for the identification of a human being comprising:
a. use of a fingerprint image (126);
b. use of two orthogonal sets of features, Minutiae (172) and Meniscus or
Global (174);
c. a matching scores fusion (180) algorithm that fuses the rank matching
scores of said Minutiae (172) with matching score of said Meniscus or
Global (174) and possibly other orthogonal feature sets (176).
7. An apparatus for transmitting data on a serial RS232 port of a computer
(106) over longer distance, comprising:
a. an enable timer (250);
b. an LED driver (252);
c. LEDs (254);
d. an RS232 interface (256);
e. An RS485 interface (258); and
f. means for providing bi-directional inexpensive conversion of a RS485
signal line 148 to RS232 signal line.
In summary, this invention is a method and apparatus that, in conjunction
with the new use of an oil impregnated material, allows the accurate
identification (verification of ID) of a human being from the fingerprints
of the said human being. The apparatus consists of an electronic
interface, FIG. 5, that can read the fingerprint of a human being, a PIN,
information stored in a card, and verify the identity of a human being
from any combination of these three pieces of information. The finger is
conditioned first by touching or pressing against, or activating a release
mechanism, that lubricates the finger, and/or clears the finger, and/or
removes excessive moisture from the finger, passively or actively the
finger by a controlled amount of lubricant, or mixture of lubricant with
disinfectant and/or fragrant.
An implementation of such a conditioning apparatus is described in our
invention, whereby the oil impregnated material MICROPOLY, manufactured by
PhyMet, Inc., of Ohio, or any variant of it that includes lubricant, oil,
fragrant, alcohol, and disinfectant in any combination, is used to
condition the finger by either touching on or pressing against its surface
either with the finger or by mechanically by a mechanism or another human
being. After conditioning the finger of a human being is tested for
quality to guarantee an average pixel intensity level above a certain
threshold. This test guarantees that the quality of the fingerprint image
is such that the false rejection probability is low. Low false rejection
probability is critical to the success of any system that relies on
fingerprints to verify the identity of a human being.
During the scanning process, the finger is calibrated electronically to
ensure consistent position of the finger on the scanner's field of view.
Audio and visual feedback is provided to enable the human being to adjust
the finger on the scanner to the correct position. A fingerprint image is
captured only when the finger is correctly positioned, thus further
reducing the probability of false rejection. The captured fingerprint
image is then electronically processed and encoded by extracting one or
more sets of features from it. These features are matched against
templates that contain similar features from the fingerprint that were
recorder previously. Each set of features is matched against a template
from similar features and a similarity score is obtained this way for each
feature set. The natural variability and wear of a finger may compromise
the ability of one particular feature set to accurately discriminate
between the correct match and an impostor.
To avoid this degradation in the identification of a human being from
fingerprint(s), multiple features sets are fused together by combining
their individual scores into a composite score, thus further reducing the
probability of false rejection and minimizing the probability of false
acceptance. In order to benefit from fusion, the fused feature sets must
be orthogonal to each other, in the sense that each one must provide
different and complimentary information from the others. Moreover, the
choice of orthogonal feature sets must be such that, that when the quality
of the information in one set is compromised due to the condition of the
finger, the information in the other feature set or sets is maintains its
integrity and can be used to identify the human being accurately. The two
feature sets in our invention, namely the Minutiae and Meniscus or Global,
meet the two criteria of (a orthogonality, and (b) complementarity.
The use of the conditioning material in conjunction with the use of fusion
of scores from multiple (more than one) orthogonal feature sets allows one
to achieve negligible false acceptance rate while maintaining a very low
false rejection rate. The combination of negligible false acceptance rate
with low false rejection rate alleviates practically all problems that
emerge from worn-our fingers and mainstreams the use of fingerprints as a
reliable and consistent means for the identification of human beings. The
new use of material can also be used as a substitute for ink and messy gel
lubricant used now in fingerprinting by law enforcement. The invention
includes a new apparatus that allows one to implement this method of
identification of a human being from fingerprints and integrate a
fingerprint scanner with alphanumeric keypad and/or electronic card reader
of any type, and the interconnection of multiple such apparatuses in a
network through a serial (daisy-chain) connection or a local area network.
The invention includes a video relay that allows one to daisy-chain in a
network multiple interface boards that are used in the reading process of
fingerprints, PINs, and electronic information on any type of a card, and
an invented new design of a bi-directional RS232-to-RS485 converter for
long-distance connections. This circuit consists of a one-shot timer that
enables the transmission of data of the RS485 network for a set time
period from the last "mark" bit sent out by the host computer. This
automatic enabling feature is necessary for the host software to work
under Windows. (Windows is a trade mark of Microsoft Corporation.) That is
because conventional RS485 converters on the market require the
Request-To-Send (RTS) line to be controlled by the host software to enable
transmission on the network. Windows will not allow a program to control
the RTS line. Automatic versions are available but are very expensive
because of the digital timing circuitry required. Our invention uses an
inexpensive analog timing circuit with loose timing tolerances for the
special case where the network protocol does not require disabling of the
transmitter the instant the last bit is sent. The User Interface Adapter
(UIA) has been designed with such a protocol. The Enable Timer circuit is
implemented with half of a very inexpensive comparator Integrated Circuit
(IC). The total number of chips required for the entire converter in only
three, as shown in FIG. 10.
DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic perspective view of the invented finger conditioning
method with oil impregnated material;
FIG. 2 is a schematic perspective view of a prior art fingerprint
conditioning process using messy lubricant gel as done by prior art;
FIG. 3 is a schematic perspective view of prior art showing the poor
quality fingerprint image captured by a scanner when the scanned finger is
dry, chapped, or subject to natural wear without the use of conditioner
(lubricant);
FIG. 4 is a schematic perspective view showing the superior quality
fingerprint image of the same dry, chapped, abraded or worn out finger of
FIG. 3, when the finger is first conditioned (lubricated) with the new oil
impregnated material prior to being scanned;
FIG. 5 is a schematic perspective view of the integrated fingerprint
scanner, keypad, card reader access control with the oil impregnated
conditioning material apparatus;
FIG. 6 is a schematic perspective view of a network of interconnected
apparatuses for human being identification that can integrate a
fingerprint scanner with the oil impregnated conditioning material,
keypad, and card reader in any possible combination on the same network;
FIG. 7 is a perspective schematic view of different packaging of the oil
impregnated material for finger conditioning for various typical usage's;
FIG. 8 illustrates a set of orthogonal features extracted from a
fingerprint, namely the Minutiae (local) features, and ridge frequency
(Meniscus or Global) features, and possibly others, fused together as
described in this invention in order to enhance the identification of a
human being;
FIG. 9 illustrates comparative performance curves of the false rejection
versus the false acceptance probabilities when Minutiae (local) only
features are used, when ridge frequency (Meniscus or Global) features are
used, and when both the orthogonal of Minutiae (local) and ridge frequency
(Meniscus or Global) information are fused together;
FIG. 10 illustrates a block diagram implementation of the invented
apparatus that is capable of interfacing information from a fingerprint
scanner, keypad, and card read in any possible combination and
interconnect multiple such apparatuses in a network by either
daisy-chaining or via a local area network;
FIG. 11 illustrates a block diagram implementation of the invented
inexpensive bi-directional RS485-to-RS232 converter for long-distance
networking of the invented apparatuses for identification of a human being
.
REFERENCE NUMERALS IN DRAWINGS
100 Finger of a human being
102 Oil impregnated material
103 Conditioned finger
104 Scanner
106 Computer
108 Set of cables
110 High quality unfragmented fingerprint image
112 Clean finger
114 Lubrication surface
115 Lubricant gel
116 Lubricant gel drops and residues
118 Gel lubricated finger
119 Wet finger
120 Low quality, smudged fingerprint image
121 Washing facility
122 Drying facility
123 Dry, chapped, or dirty finger
124 Poor quality, fragmented fingerprint
125 Conditioned dry, chapped, or dirty finger
126 High quality, unfragmented fingerprint image
128 Integrated apparatus for human being identification
130 Card reader
132 Oil impregnated MICROPOLY, or equivalent conditioning material
134 Display
136 Alphanumeric keypad
137 Local area network (LAN) cable
138 Integrated apparatus for human being identification with fingerprint
scanning and alphanumeric keypad
140 An integrated apparatus for human being identification with
alphanumeric keypad
142 An apparatus for human being identification with fingerprint scanning
and card reader
144 Integrated apparatus for human being identification with card reader
146 Integrated apparatus for human being identification with alphanumeric
keypad and card reader
148 RS485 communication protocol cable
150 Video cable
158 Oil impregnated material lubricant pad
160 Pellet
162 Screw-shaped pellet
164 Mechanical or electromechanical lubricant dispenser
166 Spring
168 Adjustable height and/or pressure receptacle
170 A Fingerprint
172 Minutiae
174 Ridge frequency information (referred to as Meniscus or Global
features)
176 Other orthogonal features (optional)
178 Feature matching scheme
180 Matching score fusion scheme
181 Fused matching scores
182 Matching characteristic with Global features only
184 Matching characteristic from Minutiae features only
186 Matching characteristic with fused Minutiae and Global features.
188 Matching accuracy characteristics
200 Microcontroller
202 Microcontroller bus
204 Address latch
206 Address decoding logic
208 ROM (Read Only Memory)
210 Additional ROM
212 Switch Interface
214 Numeric Keypad
216 Alphabetic Keypad
218 Configuration switches
220 Control register
222 Door latch relay driver
224 Video relay driver
226 Cover motor relay driver
228 LCD (Light Coupled Display) backlight driver
230 Backlit LCD display
232 CPU expansion port
234 Audio amplifier
236 Speaker
238 RS232 network interface
240 RS485 network interface
242 Card reader
244 External switch interface
246 Door latch relay
248 Video relay
250 Enable timer
252 LED (Light Emitting Diode) driver
254 LEDs
256 RS232 Interface
258 RS485 Interface
DETAILED DESCRIPTION OF THE INVENTION AND BEST MODE FOR CARRYING OUT THE
PREFERRED EMBODIMENT
Referring to FIG. 1 a typical fingerprint image capturing process according
to this invention is described whereby a finger of a human being 100 is
first conditioned using an oil impregnated material 102. Finger
conditioning is achieved by pressing or rotating the finger 100 against
the oil impregnated material 102 to provoke the release of a controlled
amount of the lubricant from the oil impregnated material 102. The
lubricant can be any oily substance, or any mixture of oily substance with
fragrant or disinfectant or both. The finger 100 after conditioning 103 is
placed on a scanner 104 and a high quality unfragmented fingerprint image
110 is captured by a computer 106 which is connected to the scanner via a
set of cables 108. Conditioning of the finger 100 with a mixture of a
lubricant, fragrant and disinfectant achieves: (a) the lubrication that is
required to condition the finger 100 against dryness, chapping, dirt, or
wear-out so as to guarantee the quality fingerprint image 110; (b)
deodorizes the finger with a pleasant sent; and (c) disinfects both the
finger 100 and the scanner 104 after each use and makes the use of our
invention contamination-free, safe to use. The high quality fingerprint
image 110 is then processed and matched against one or more candidate
templates from a database to identify the human being. The human being
leaves the fingerprint identification process with a clean finger 112
without the need of washing or drying the finger 100 after conditioning
103.
Referring to FIG. 2, a typical cumbersome and messy finger conditioning
process according to prior art is described, whereby the finger 100 is
first deepen in and covered with a light and/or sound conducting lubricant
gel 115 on a lubrication surface 114 with lubricant gel 115. A gel
lubricated and gel dripping finger 118 is then placed on the scanner 104
and a low quality, smudged fingerprint image 120 is captured by the
computer 106 which is connected to the scanner through the set of cables
108. The low quality, smudged fingerprint image is processed and matched
against one or more templates from a database with questionable accuracy
due to the low quality and smudges. At the end of the fingerprint
identification process, the gel lubricated finger 118 requires washing
with soap and water in a washing facility 121 and the wet finger 119
requires drying with towel (cloth or paper) in a drying facility 122.
Moreover, the gel conditioning process has left messy lubricant gel drops
and residue 116 on the lubrication surface 114, the scanner 104, and the
gel lubricated finger 118 before it is cleaned off.
Referring to FIG. 3, a typical fingerprint scanning process of a dry,
chapped, or dirty finger 123 according to prior art, wherein the dry,
chapped, or dirty finger 123 is placed on the scanner 104 without prior
conditioning. The poor quality, fragmented fingerprint image 124 which is
captured by the computer 124 which is connected to the scanner through the
cable 108 is unacceptable for the accurate identification of
identification of a human being.
Referring to FIG. 4, according to our invention, when the same dry,
chapped, or dirty finger 123 is first conditioned by putting it in contact
with the oil impregnated material 102 and a dry, chapped, or dirty finger
after conditioning 125 is placed on the scanner 104, a high quality
unfragmented fingerprint image 126 is captured by the computer 106 which
connects to the scanner 104 with the cable set 108. The characteristic
features that make each fingerprint unique (and are described further down
in our invention) can be extracted easily and unambiguously from the high
quality, unfragmented fingerprint image 126 to guarantee the accurate
identification of a human being. The use of the oil impregnated material
104, along with the use of multiple orthogonal feature sets in fusion,
allows one to build an accurate, dependable, and consistent system for the
identification of a human being suitable for unattended applications, such
as access control, intrusion control, positive verification for financial
transactions, as well as attended applications such as fingerprinting of
suspects, personal verification for financial aid recipients, and criminal
identification.
Referring to FIG. 5, an apparatus according to this invention is described
whereby the oil impregnated material 102 is packaged in a pellet 132 and
attached to an integrated man-machine interface 128 that is used in the
process of identifying a human being from fingerprints. The interface 128
consists of the fingerprint scanner 104, an LCD display 134, an
alphanumeric keypad 136, and a card reader 130. This interface connects to
an electronic processor, either external or internal, and is used to
provide accurate and unambiguous identity verification of a human being. A
human being first conditions his/her finger by pressing and swiping the
finger lightly against the oil impregnated pellet 132, then punching a
Personal Identification Number (PIN) on the keypad or swiping or inserting
a card in the card reader, or using both in combination to provide a
unique ID. The integrated apparatus 128 verifies that the PIN and/or the
card number are constitute a valid ID, and signals the human being to
place the finger on the scanner 104. The integrated interface 128 measures
the position and quality of the fingerprint image and uses visual feedback
on the display 134 and audio feedback through a speaker to notify the
human being to: (a) adjust the finger on the scanner 104 when the finger
is not properly placed, and (b) to repeat the lubrication process if
needed.
The finger 100 is determined to be positioned high or low on the scanner
104, by scanning across the finger 100 at three row locations of the
fingerprint image 110. Fingerprint ridges show up as low intensity pixels
against the white background intensity of the scanner 104. If the ratio of
the number of low intensity pixels at a row next the top of the
fingerprint image, to the number of low intensity pixels at the middle of
the fingerprint image (110) is greater than a selected threshold, the
fingerprint is considered to be too high, and requires re-positioning the
finger 110 on the scanner 104. Determination of the quality and stability
of the fingerprint image (110) prior to capturing the fingerprint image
(110) is accomplished by sampling the fingerprint image at three
sequential time intervals and by inspecting the relative changes in the
fingerprint image 110. A fingerprint image 110 is considered suitable for
acceptance and testing of the identity of said human being, if three
conditions are met. The first condition is that sufficient surface of the
fingerprint image 110 has been captured by the scanner 104. This is
measured by comparing low intensity (non-background) pixels on the screen
to a minimum threshold value indicative of a minimum surface area. If the
number of pixels is greater than the threshold, this condition is
considered to be true.
| | |
