“Biometric technologies” are automated methods of verifying or recognizing the identity of a living person based on a physiological or behavioral characteristic. Our research includes fingerprint, iris and face based biometrics and related works such as image quality assessment, liveness detection and performance evaluation.

Biometric researches performed in our laboratory can be divided into 3 categories:


Face 3D Scan using Multi-Kinect system

Nowadays, face recognition in video surveillance systems is a challenging task, because it is very difficult to match the frontal 2D face and the rotated 2D faces. A solution to the face rotation problem is utilizing 3D faces. However, current commercial 3D scanner products are very expensive. In this work, we proposed a 3D face scanning system using multiple Microsoft Kinects, which is affordable for consumers and provides good performance.

System Environment (3 Kinect)

3D face data

Data acqusition process & experiment result


Related Papers:

    • 김정민, 이성철, 부성채, 김주성, 김학일, Seong G. Kong : Multi-Kinect를 이용한 3D 얼굴 모델링 및 2D-3D 얼굴인식 성능평가. IPIU 2013

Back to top


Level of Difficulty of Fingerprint Databases

Understanding the difficulty of a dataset is of primary importance when it comes to testing and evaluating fingerprint recognition systems or algorithms because the evaluation result is dependent on the dataset. Our propoed method is a general framework of assessing the level of difficulty of fingerprint datasets based on quantitative measurements of not only the sample quality of individual fingerprints but also relative differences between genuine pairs, such as common area and deformation.

Example of measuring common area and deformation of mated pair
Level of Difficulty vs EER

Related papers:

    • Changlong Jin, Shengzhe Li, Hakil Kim, “Type-independent pixel-level alignment point detection for fingerprints”, 2011 International Conference on Hand-Based Biometrics (ICHB), Hongkong
    • Shengzhe Li; Changlong Jin; Hakil Kim; Elliott, S., “Assessing the Difficulty Level of Fingerprint Datasets Based on Relative Quality Measures “, 2011 International Conference on Hand-Based Biometrics (ICHB), Hongkong

Back to top

Fingerprint Pre-processing

A reliable pixel-wise orientation field (OF) estimation method is of primary importance in fingerprint pattern recognition. Robust high resolution OF not only enhances the accuracy in position of singular points but also reduces the number of false singular points.

Pixel-wise orientation field (OF) estimation.
Candidate singular point (SP) detection

Related papers:

    • Changlong Jin, Hakil Kim: Pixel-level singular point detection from multi-scale Gaussian filtered orientation field. Pattern Recognition 43(11): 3879-3890 (2010)

Back to top

Fingerprint Liveness Detection and Image Quality

In spite its advantages, fingerprint systems are vulnerable to attacks. Many studies showed that artificial fingers made by silicone, gelatin or other materials can spoof most of the existing fingerprint sensors including optical sensor, capacitive sensor, electronic sensor, etc. Making a fake finger and presenting it to the sensor is the most common and easy way to spoof fingerprint system since it is not so hard to get the fake materials and easy to make the fake fingertip.

Fake fingerprint made of silicon

Fingerprint liveness detection can be performed by checking the quality differences such as the fingerprint integrity, middle ridge line signals, power spectrum feature, gray-level values, contour shape of foreground, etc. Although the quality-based approach cannot distinguish the fake fingerprints with high quality from the live fingerprints in identification mode, it can find out the quality difference in verification mode. Namely, if the quality difference between the enrolled sample and the test sample is large, then the test sample may be a fake fingerprint.

Classification of fake and live fingerprints
Classification results of fingerprint liveness detection

Related papers:

    • Changlong Jin, Shengzhe Li, Hakil Kim, Eunsoo Park: Fingerprint Liveness Detection Based on Multiple Image Quality Features. WISA 2010: 281-291
    • Changlong Jin, Hakil Kim, Stephen J. Elliott: Liveness Detection of Fingerprint Based on Band-Selective Fourier Spectrum. ICISC 2007: 168-179

Back to top


Iris Recognition and Eye Detection

Eye detection plays an important role in applications related to face recognition. The position of eyes can be used as a reliable reference for other facial feature detection. Our method adopts a novel approach for the precise and reliable detection of eyes by introducing a ternary eye-verifier. The method achieves precise and reliable detection of eyes from color facial images with variation in illumination, pose, eye gazing direction, and race.

Example of eye detection

Related Papers:

    • Van Huan Nguyen, Thi Hai Binh Nguyen, Hakil Kim: Eye feature extraction using K-means clustering for low illumination and iris color variety. ICARCV 2010: 633-637
    • Van Huan Nguyen, Hakil Kim: Robust iris segmentation via simple circular and linear filters. J. Electronic Imaging 17(4): 043027 (2008)
    • Van Huan Nguyen, Thi Hai Binh Nguyen, Hakil Kim: Location of iris based on circular and linear filters. ICARCV 2008: 412-416

Back to top


Facial image quality

ISO/IEC 19794-5 defines a standard scheme for both image quality requirements as well as the data formats to be a CBEFF (Common Biometric Exchange Formats Framework)-compliant for facial sample data. This research aims at developing a computerized conformance testing mechanism based on this standard to automatically assure a facial passport photo to be ISO/IEC 19794-5-compliant. Our research is to investigate the requirements for image quality, namely, scene requirements (such as lighting, pose, expression, etc.); photographic requirements (positioning, camera focus, etc.); digital requirements (image resolution, image size, etc.), but data format structure (an aspect of data storing, transferring).

Face image quality

Related papers:

    • Thi Hai Binh Nguyen, Van Huan Nguyen, Hakil Kim: Combination of edge and color information for robust preprocessing in facial image quality assessment. SMC 2010: 3594-3600
    • Thi Hai Binh Nguyen, Van Huan Nguyen, Hakil Kim: Robust Feature Extraction for Facial Image Quality Assessment. WISA 2010: 292-306

Back to top