|
Biometrics
Robotics
Resources
Ordering
Services
|
MegaMatcher Algorithm Features and Capabilities
Performance numbers are provided for a PC with Intel Core 2 Q9400 processor (2.67 GHz).
Download
MegaMatcher SDK
brochure (PDF)
Complete information, including all technical specifications, licensing and prices.
The 53-page brochure can be printed on both Letter and A4 paper.
File size:
2.6 Megabytes;
Updated on:
April 13, 2012.
MegaMatcher includes fingerprint, facial, speaker, iris and palm print recognition engines along with a fused algorithm for fast and reliable identification in large-scale systems.
The fingerprint, face, voice and iris identification algorithms may each be used separately to develop AFIS, automated face, speaker or iris identification systems.
The biometric software engines contain many proprietary algorithmic solutions that are especially useful for large-scale identification problems.
These solutions were specifically developed for MegaMatcher, incorporating aspects of the VeriFinger, VeriLook, VeriSpeak and VeriEye algorithms.
Some of these solutions are listed in the fingerprint, face, voice and iris biometric identification engine descriptions below.
MegaMatcher fingerprint template extraction and matching engine
-
Full MINEX Compliance.
NIST has recognized MegaMatcher fingerprint algorithm as MINEX compliant and suitable for use in personal identity verification (PIV) program applications.
-
Rolled and flat fingerprints matching.
The MegaMatcher fingerprint engine matches rolled and flat fingerprints between themselves.
Typically, conventional "flat" fingerprint identification algorithms perform matching between flat and rolled fingerprints less reliably due to the specific deformations of rolled fingerprints.
MegaMatcher allows flat-to-flat, flat-to-rolled or rolled-to-rolled fingerprint matching with a high degree of reliability and accuracy.
The algorithm matches up to 136,000 flat fingerprint records per second.
-
MegaMatcher includes fingerprint image quality determination, which may be used during enrollment to ensure that only the best quality fingerprint template will be stored in the database.
-
Template generalization is used to generate a better quality template from several fingerprints.
Better quality templates result in a higher level of identification accuracy.
-
MegaMatcher is tolerant to fingerprint translation, rotation and deformation.
It uses a proprietary fingerprint matching algorithm that identifies fingerprints even if they are rotated, translated or have deformations.
-
Adaptive image filtration algorithm eliminates noises, ridge ruptures and stuck ridges, and reliably extracting minutiae from even the poorest quality fingerprints in less than 1 second.
MegaMatcher face template extraction and matching engine
-
Template generalization is used to generate a better quality template from several face images.
Better quality templates result in a higher level of identification accuracy.
-
Tolerance to face position assures a level of enrollment convenience.
MegaMatcher allows for 360 degrees of head roll.
Head pitch can be up to 15 degrees in each direction from the frontal position.
Head yaw can be up to 45 degrees in each direction from the frontal position.
See technical specifications for more details.
-
Reliable face detection assures accurate enrollment from cameras, webcams and various scanned documents; faces may be enrolled from the scanned pages of passports or other types of documentation.
When there are multiple faces present in a video or an image, they may be enrolled and processed simultaneously.
-
Live face detection.
A conventional face identification system can be tricked by placing a photo in front of the camera.
MegaMatcher is able to prevent this kind of security breach by determining whether a face in a video stream is "live" or a photograph.
-
The biometric template record can contain several face samples belonging to the same person.
These samples can be enrolled from different sources and at different times, thus allowing improvement in matching quality.
For example a person might be enrolled with eyeglasses and without, or with different types of eyeglasses; with and without beard or moustache, etc.
MegaMatcher voice template extraction and matching engine
-
Text-dependent voice matching engine determines if a voice sample matches the template that was extracted from a specific phrase.
During enrollment, one or more phrases are requested from the person being enrolled.
Later that person may be asked to pronounce a specific phrase for verification.
This method assures protection against the use of a covertly recorded random phrase from that person.
-
Two-factor authentication with a passphrase is performed when a person is asked to say a unique phrase (such as passphrase or an answer to a "secret question" that is known only by the person being enrolled).
The overall system security increases as both voice authenticity and password are checked.
-
Liveness detection.
A system may request each user to enroll a set of unique phrases.
Later the user will be requested to say a specifc phrase from the enrolled set.
This way the system can ensure that a live person is being verified (as opposed to impostor who uses voice recording).
-
Several voice records with the same phrase may be stored to improve speaker recognition reliability.
Certain natural voice variations (i.e. hoarse voice) or environment changes (i.e. office and outdoors) can be stored in the same template.
MegaMatcher iris template extraction and matching engine
-
NIST IREX proven reliability.
MegaMatcher iris matching engine is based on VeriEye, that was recognized by NIST in 2009 as one of the most reliably accurate iris recognition algorithms.
-
Fast matching.
Configurable matching speed varies from 60,000 to 920,000 comparisons per second.
See technical specifications for more details.
-
Robust iris detection.
Irises are detected even when the images have obstructions, visual noise and different levels of illumination.
Lighting reflections, eyelids and eyelashes obstructions are eliminated.
Images with narrowed eyelids or eyes that are gazing away are also accepted.
-
Automatic interlacing detection and correction.
The correction results in maximum quality of iris features templates from moving iris images.
-
Correct iris segmentation is achieved when perfect circles fail, the centers of the iris inner and outer boundaries are different, iris boundaries are definitely not circles and even not ellipses or iris boundaries seem to be perfect circles.
Go to MegaMatcher contents
Technical Specifications
All biometric templates should be loaded into RAM before identification, thus the maximum biometric templates database size is limited by the amount of available RAM.
-
Fingerprint scanners are recommended to have at least 500 ppi resolution and at least 1" x 1" fingerprint sensors.
The specifications are provided for 500 x 500 pixels fingerprint images and templates extracted from these images.
-
Face capture cameras are recommended to produce at least 640 x 480 pixels images for reliable faces' detection.
Face template extraction and matching speed is not dependent on the image size.
-
The minimal distance between eyes is 50 pixels for a face on image or video stream to perform face template extraction.
75 pixels or more recommended for better template extraction results.
-
Face recognition engine has certain tolerance to face posture:
-
head roll (tilt) – ±180 degrees (configurable);
±15 degrees default value is the fastest setting which is usually sufficient for most near-frontal face images.
-
head pitch (nod) – ±15 degrees from frontal position.
-
head yaw (bobble) – ±45 degrees from frontal position.
±15 degrees default value is the fastest setting which is usually sufficient for most near-frontal face images.
The specifications are provided for the default roll and yaw values.
-
Iris capture cameras are recommended to produce at least 640 x 480 pixels images.
The specifications are provided for these images.
-
At least 2-second long voice samples are recommended to assure speaker recognition quality.
-
At least 11,025 Hz sampling rate with at least 16-bit depth should be configured during voice recording.
See also the whole lists of recommendations and constraints for facial recognition and speaker recognition.
MegaMatcher biometric template matching algorithm can be run on more than one processor core on multi-core processors allowing to increase template matching speed.
The template matching speeds in the table below are given as a range, where the smaller number means matching speed using 1 processor core, while the larger number means matching speed using 4 processor cores.
The specifications are provided for these processors:
- Intel Core 2 Q9400 (4 cores), running at 2.67 GHz clock rate;
- Intel Core i7-2600 (4 cores), running at 3.4 GHz clock rate.
| MegaMatcher 4.3 fingerprint template extraction and matching engine specifications |
| |
Intel Core 2 Q9400 |
Intel Core i7-2600 |
Maximized
matching
accuracy |
Maximized
matching
speed |
Maximized
matching
accuracy |
Maximized
matching
speed |
| Template extraction time (seconds) |
0.20 - 0.25 |
0.12 - 0.15 |
Template matching speed(1) with
±45° fingerprint rotation tolerance
(fingerprints per second) |
19,000 - 76,000 |
34,000 - 136,000 |
38,500 - 154,000 |
70,000 - 280,000 |
Template matching speed(1) with
±90° fingerprint rotation tolerance
(fingerprints per second) |
15,000 - 60,000 |
27,000 - 108,000 |
31,000 - 124,000 |
55,000 - 220,000 |
Template matching speed(1) with
±180° fingerprint rotation tolerance
(fingerprints per second) |
11,000 - 44,000 |
20,000 - 80,000 |
22,500 - 90,000 |
41,000 - 164,000 |
| Single fingerprint record size in a template(2) (bytes) |
700 - 6,000
(configurable) |
| MegaMatcher 4.3 face template extraction and matching engine specifications |
| |
Intel Core 2 Q9400 |
Intel Core i7-2600 |
Maximized
matching
accuracy |
Maximized
matching
speed |
Maximized
matching
accuracy |
Maximized
matching
speed |
| Template extraction time (seconds) |
0.19 |
0.07 |
0.11 |
0.04 |
Template matching speed(1)
(faces per second) |
12,000 - 48,000 |
215,000 - 860,000 |
30,000 - 120,000 |
525,000 - 2,100,000 |
| Single face record size in a template(2) (bytes) |
35,994 |
4,026 |
35,994 |
4,026 |
| MegaMatcher 4.3 iris template extraction and matching engine specifications |
| |
Intel Core 2 Q9400 |
Intel Core i7-2600 |
Maximized
matching
accuracy |
Maximized
matching
speed |
Maximized
matching
accuracy |
Maximized
matching
speed |
| Template extraction time (seconds) |
0.10 - 0.12 |
0.08 - 0.10 |
Template matching speed(1) with
±15° iris rotation tolerance
(irises per second) |
60,000 - 240,000 |
230,000 - 920,000 |
130,000 - 520,000 |
550,000 - 2,200,000 |
Template matching speed(1) with
±30° iris rotation tolerance
(irises per second) |
35,000 - 140,000 |
175,000 - 700,000 |
70,000 - 280,000 |
350,000 - 1,400,000 |
| Single iris record size in a template(2) (bytes) |
2,328 |
| MegaMatcher 4.3 voiceprint template extraction and matching engine specifications(3) |
| |
Intel Core 2 Q9400 |
Intel Core i7-2600 |
Fixed
phrase |
Unique
phrase |
Fixed
phrase |
Unique
phrase |
| Template extraction time (seconds) |
0.12 - 0.15 |
0.08 - 0.10 |
Template matching speed(1)
(voiceprints per second) |
230 - 920 |
140 - 560 |
450 - 1,800 |
250 - 1,000 |
| Single voiceprint record size in a template(2) (bytes) |
4,500 - 5,000 |
Notes:
(1) The speeds are given for a single PC with the specified processor. If a cluster is used, the speeds should be multiplied by the number of cluster nodes.
(2) MegaMatcher 4.3 allows to store multiple biometric records of the same or different biometric modalities in a template; in this case the template size is the sum of all included biometric records.
(3) The specifications are provided for 5-second long voice samples; template size and template extraction speed have linear dependence from voice sample length, template matching speed has quadratic dependence from voice sample lenght.
See also: technical specifications for MegaMatcher Palm Print engine.
Go to MegaMatcher contents
Reliability and Performance Testing Results
The identification reliability and speed are important for large-scale systems.
MegaMatcher SDK includes a fused algorithm for fast and reliable identification using several biometric records taken from the same person.
As we do not have any single database with all supported biometric modalities, separate tests with selected modalities were performed for the MegaMatcher biometric engines to demonstrate their reliability and performance with single biometric modalities and combinations of several modalities:
Single modality engines tests
These reliability tests show the MegaMatcher 4.3 fingerprint, face, iris and voiceprint single modality engines performance and reliability.
The tests were performed with these publicly available databases:
The biometric engines had these parameters set:
-
±90 degrees fingerprint rotation tolerance value was used for template matching;
-
±15 degrees iris rotation tolerance value was used for template matching.
These tests were performed with the biometric databases:
-
Test 1 maximized matching accuracy.
The test was performed with all databases.
MegaMatcher 4.3 fused algorithm reliability in this test is shown as blue curves on the ROC charts.
-
Test 2 maximized matching speed.
The test was performed only with fingerprint, face and iris databases.
MegaMatcher 4.3 fused algorithm reliability in this test is shown as red curves on the ROC charts.
Receiver operation characteristics (ROC) curves are usually used to demonstrate the recognition quality of an algorithm.
ROC curves show the dependence of false rejection rate (FRR) on the false acceptance rate (FAR).
| MegaMatcher 4.3 template matching engines reliability testing results |
| A template contains these biometric records |
FRR at
0.001 % FAR |
FRR at
0.0001 % FAR |
| Test 1 |
Test 2 |
Test 1 |
Test 2 |
| 1 fingerprint |
0.230 % |
0.294 % |
0.280 % |
0.384 % |
| 1 face |
2.363 % |
3.647 % |
3.611 % |
4.750 % |
| 1 iris |
1.772 % |
1.834 % |
2.270 % |
2.352 % |
| 1 voiceprint |
33.820 % |
43.930 % |
Template matching was performed using all 4 cores of the specified processors.
The performance tests were performed on PCs with these processors:
- Intel Core 2 Q9400, running at 2.67 GHz clock rate;
- Intel Core i7-2600, running at 3.4 GHz clock rate.
| MegaMatcher 4.3 single modality matching engines speed testing results (templates per second) |
| A template contains these biometric records |
Core 2 Q9400 |
Core i7-2600 |
| Test 1 |
Test 2 |
Test 1 |
Test 2 |
| 1 fingerprint |
61988 |
108136 |
124784 |
220232 |
| 1 face |
48736 |
869248 |
115072 |
2187344 |
| 1 iris |
245212 |
920692 |
526376 |
2205376 |
| 1 voiceprint |
532 |
912 |
Fingerprint, face and iris matching engines tests
The tests with MegaMatcher biometric fingerprint, face and iris matching engines and fused template matching algorithm were performed using Neurotechnology internal multi-biometric database:
- The database had 7,500 sets of biometric records; each set contained 1 face, 2 irises and 10 fingerprints representing a unique person.
- 1,500 unique persons were represented in the database.
- 5 capture sessions were performed for each person.
The tests were performed with these biometric template types:
- 1 fingerprint record extracted from left index fingerprint image.
- 1 face record.
- 1 iris record extracted from left eye image.
- 2 fingerprint records extracted from same person's left and right index fingerprint images.
- 2 iris records extracted from same person's different eye images.
- 1 fingerprint + 1 face records – left index fingerprint and face taken from the same person.
- 1 face + 1 iris records – left iris and face taken from the same person.
- 1 fingerprint + 1 iris records – left index fingerprint and left iris taken from the same person.
- 1 fingerprint + 1 face + 1 iris records – left index fingerprint, left iris and face taken from the same person.
The biometric engines had these parameters set:
-
±90 degrees fingerprint rotation tolerance value was used for template matching;
-
±15 degrees iris rotation tolerance value was used for template matching.
Two tests were performed with each template type:
-
Test 1 maximized matching accuracy.
MegaMatcher 4.3 fused algorithm reliability in this test is shown as blue curves on the ROC charts.
-
Test 2 maximized matching speed.
MegaMatcher 4.3 fused algorithm reliability in this test is shown as red curves on the ROC charts.
Receiver operation characteristics (ROC) curves are usually used to demonstrate the recognition quality of an algorithm.
ROC curves show the dependence of false rejection rate (FRR) on the false acceptance rate (FAR).
| MegaMatcher 4.3 template matching engines reliability testing results |
| A template contains these biometric records |
FRR at
0.001 % FAR |
FRR at
0.0001 % FAR |
| Test 1 |
Test 2 |
Test 1 |
Test 2 |
| 1 fingerprint |
0.230 % |
0.280 % |
0.294 % |
0.384 % |
| 1 face |
12.680 % |
15.790 % |
18.470 % |
20.910 % |
| 1 iris |
1.167 % |
1.253 % |
1.467 % |
1.547 % |
| 2 fingerprints |
0.047 % |
0.054 % |
0.047 % |
0.074 % |
| 2 irises |
0.293 % |
0.330 % |
0.323 % |
0.397 % |
| 1 fingerprint + 1 face |
0.020 % |
0.030 % |
0.020 % |
0.047 % |
| 1 fingerprint + 1 iris |
0.000 % |
0.000 % |
0.000 % |
0.000 % |
| 1 face + 1 iris |
0.203 % |
0.333 % |
0.323 % |
0.443 % |
| 1 fingerprint + 1 face + 1 iris |
0.000 % |
0.000 % |
0.000 % |
0.000 % |
Template matching was performed using all 4 cores of the specified processors.
The performance tests were performed on PCs with these processors:
- Intel Core 2 Q9400, running at 2.67 GHz clock rate;
- Intel Core i7-2600, running at 3.4 GHz clock rate.
| MegaMatcher 4.3 matching engines speed testing results (templates per second) |
| A template contains these biometric records |
Core 2 Q9400 |
Core i7-2600 |
| Test 1 |
Test 2 |
Test 1 |
Test 2 |
| 1 fingerprint |
61998 |
108136 |
124784 |
220232 |
| 1 face |
62560 |
1037720 |
133504 |
2441736 |
| 1 iris |
253476 |
1032628 |
542840 |
2424296 |
| 2 fingerprints |
30452 |
54176 |
60992 |
109528 |
| 2 irises |
131568 |
592160 |
279288 |
1437592 |
| 1 fingerprint + 1 face |
29784 |
97760 |
63856 |
201096 |
| 1 fingerprint + 1 iris |
49300 |
97820 |
100792 |
201336 |
| 1 face + 1 iris |
49924 |
554180 |
106992 |
1334264 |
| 1 fingerprint + 1 face + 1 iris |
26508 |
89300 |
56824 |
184136 |
These tests show that a large-scale automated biometric identification system based on MegaMatcher provides high identification reliability when using fingerprints, using fused same-biometric (different fingerprints or irises from the same person) matching significantly reduces FRR, and using multi-biometric identification results in a significant reliability increase, allowing the system to reach almost 0 % FRR.
Voiceprint and face matching engines tests
The tests with MegaMatcher biometric face and voiceprint matching engines, and the fused template matching algorithm were performed using face images and voice samples from the XM2VTS Database:
- 295 unique persons were represented in the database.
- 8 capture sessions were performed for each person.
- The phrase 1 from the database was used for the testing, meaning that the same fixed phrase was used for all subjects.
The tests were performed with these biometric template types:
- 1 voiceprint record.
- 1 face record.
- 1 voiceprint + 1 face records taken from the same person.
Receiver operation characteristics (ROC) curves are usually used to demonstrate the recognition quality of an algorithm.
ROC curves show the dependence of false rejection rate (FRR) on the false acceptance rate (FAR).
Template matching was performed using all 4 cores of the specified processors.
The performance tests were performed on PCs with these processors:
- Intel Core 2 Q9400, running at 2.67 GHz clock rate;
- Intel Core i7-2600, running at 3.4 GHz clock rate.
| MegaMatcher 4.3 face, voiceprint and fused template matching engines tests |
| A template contains these biometric records |
Matching speed
(templates per second) |
FRR at
0.001 % FAR |
FRR at
0.0001 % FAR |
| Core 2 Q9400 |
Core i7-2600 |
| 1 voiceprint |
532 |
912 |
33.820 % |
43.930 % |
| 1 face |
48736 |
115072 |
2.387 % |
3.574 % |
| 1 voiceprint + 1 face |
524 |
912 |
0.666 % |
1.017 % |
Go to MegaMatcher contents
|
Products
AFIS or multi-biometric fingerprint, iris, face and voice identification for large-scale systems.
Face identification for PC or Web solutions.
Fingerprint identification for PC and Web solutions.
Iris identification for PC and Web solutions.
Speaker recognition for PC or Web applications.
Object recognition for robotics and computer vision.
More products for developers:
End-user products:
- NCheck Finger Attendance – an attendance control application that uses fingerprint biometrics to perform employee identification.
- NVeiler Video Filter – a plug-in for VirtualDub that automatically detects faces in a frame, tracks the faces (or other objects) in subsequent frames and hides them.
|
