AMT-FaceMatch & AMT-PalmMatch

Supported by AMT-FAPVSM-21

Specifications
Biometric Type

Face

Palm

Recognition Distance

50 cm ~ 160 cm /
20 inch ~ 63 inch

15 cm ~ 40 cm /
6 inch ~ 15 inch
Posture Adaptability
(by Default)

Yaw ≤ 30°, Pitch ≤ 30°,
Roll ≤30°

Yaw≤30°, Pitch≤30°,
Roll≤180°, Bend≤20%

Accuracy

*TAR=99%@FAR=0.01%

*TAR=99.32%@FAR=0.001%

Recognition Mode

1:1, 1:N

1:1, 1:N

Template Extraction Time

< 200 ms

< 300ms

Matching Time

< 70 ms

< 150 ms

Face Attribute

Age, Gender,
Expression class

N/A

Liveness Detection

YesYes
Face Mask Detection

YesN/A

Capacity

100,000

6,000

* The performance test is based on the Armatura proprietary face and palm dataset.

AMT Facelite SDK

Supported by AMT-FAM-10

Specifications
AlgorithmAMT FaceLite Algorithm v12.0 
SDK SizeAndroid Jar Lib < 15MB
Windows DLL lib< 130MB
Pose RangeYaw ≤ 25°, Pitch ≤ 25°, Roll ≤ 25°
Template Size28KB
Performance*Face Detection Time< 50ms
Template Extraction Time< 200ms
Identification Time< 100ms
Match Mode1:1 for verification, 1:N for identification
Capacity 6,000 templates
Accuracy**TAR =98.6% when FAR = 0.001%
Operation SystemWindows Windows XP, 7, 10 (x86 & x64)
AndroidAndroid 4.1 and above
LinuxProvided on request
Demo ProgramC/C++, C#, Java
Note: * The algorithm performance is assessed based on image resolution of 480 * 640, and Quad-Core Cortex-A9, 1.5GHz processor.
** The accuracy is assessed based on the proprietary infrared light face image data set.
** TAR: True Acceptance Rate, FAR: False Acceptance Rate.

AMT PalmLite SDK

Supported by AMT-PVM-10 & AMT-PVR-10

Specifications
AlgorithmAMT PalmLite Algorithm v12.0 
SDK SizeAndroid Jar Lib < 5MB
Windows DLL lib< 6MB
Pose RangeYaw ≤ 30°, Pitch ≤ 30°, Roll ≤ 180°, Bend ≤ 20%,
Template Size8844 Bytes
Performance*Palm Detection Time< 50ms
Template Extraction Time< 200ms
Identification Time (1:6000)< 150ms
Match Mode1:1 for verification, 1:N for identification
Capacity 6,000 templates
Accuracy**TAR =99.3% when FAR = 0.001%
Operation SystemWindows Windows XP, 7, 10 (x86 & x64)
AndroidAndroid 4.1 and above
LinuxProvided on request
Demo ProgramC/C++, C#, Java
Note: *The algorithm performance is assessed based on image resolution of 480 * 640 and Quad-Core Cortex A9 CPU@1.5GHz platform.
** The accuracy is assessed based on the proprietary infrared light palm image data set.
** TAR: True Acceptance Rate, FAR: False Acceptance Rate.

Multi Bio SDK

Supported by AMT-FAPVSM & AMT-FAPVSR

Get in touch

Armatura USA Headquarters

190 Bluegrass Valley Pkwy,
Alpharetta, GA 30005

Phone : + 1 (470) 816-1970