메인컨텐츠로 바로가기
Products

A·I based Algorithm

Fingerprint Recognition Algorithm

Features

  • Proven by its successful case of stable application to e-Gate (Ministry of Justice, Korea)
  • Designed to build up the biometric system by providing various APIs
  • Fast system and stable performance for 1:1 and 1:N matching
  • Supports a variety of platforms including Windows, Linux, Unix, and Android
  • Certified by the only official biometric certification authority in Korea, KISA (Korea Internet & Security Agency) for the remarkable performance of CUBOX Fingerprint Algorithm

Features

Reliability under variable conditions
(rotation, displacement, distortion, etc.)
Outstanding recognition performance through global and local matching
로컬매칭
Powerful preprocessing algorithm
Adaptive contrast enhancement and adaptive filtering to extract accurate minutiae from wet or dry fingerprints
Rolled or flat fingerprint matching
Applicable for the systems where deformed fingerprints are to be tracked down, such as in fingerprint identification for police system
Quality measurement
NFIQ (NIST Fingerprint Image Quality) or CUBOX’s measurement function is provided to assure high image quality
Quality:24 Quality:56 Quality:92
  • 1-to-1, 1-to-many matching function: 1:N identification that is capable of comparing 200,000 fingerprints/sec, as well as 1:1 verification mode, is provided
  • ISO-19794-2,4 Standard Format support:Inter-vender compatibility due to the support for the international standard format
  • Support for various image format and conversion available to WSQ, BMP, JPG., etc.
  • Size of feature data adjustable in accordance with individual applications, resulting in efficient use of system memory.
  • Due to their properties, feature data size can be adjusted, resulting in efficient use of system memory and can also be applied to smaller devices such as IC chips.

Test Results

Testing environment and evaluation index

Testing environment
Hardware : Intel® Core™ i7-3770 3.40GHz, 8GB memory
Software : Microsoft Windows 8 K
Evaluation index
EER(Equal Error Rate)
Error rate at which FMR (False Match Ratio) and FNMR (False Not Match Ratio) coincide
FTA(Failure To Attempt)
FTE(Failure To Enrollment)

Results

Index DIG HUN PER
EER(%) 0.33 0.50 0.14
FTA(%) 0 0 0
FTE(% 0 0 0
ROC OF DIG ROC OF HUN ROC OF PER