<<HTML(<meta http-equiv="refresh" content="0;URL=../../CMV/Research/FaceRecognitionAndBiometrics">)>>
<<HTML(<a href="#" onclick="history.back(); return false;">Back</a>)>>
Face Recognition and Biometrics
Automatic face analysis has become a very active topic in computer vision research as it is useful in several applications, like biometric identification, visual surveillance, human-machine interaction, video conferencing and content-based image retrieval. Face analysis is challenging due to the fact that a face is a dynamic and non-rigid object which is difficult to handle. Its appearance varies due to changes in pose, expression, illumination and other factors such as age and make-up.
We have a long experience in investigating face related problems since 1997. First, the research was focused on colour-based skin detection. A robust method for skin colour modelling called "skin locus" was developed providing good performance in varying illumination conditions. To further analyse the properties of skin chromaticities under changing lighting conditions, the group collected and made publicly available two test image databases (Physics-Based Face Database and Face Video Database), which have been delivered to a large number of academic and industrial research groups all over the world.
In 2004, the group introduced a novel texture-based approach for face description. The approach is based on the efficient local binary pattern (LBP) features developed by the group a few years ago. Excellent results were obtained in face detection, face recognition and facial expression recognition. Several other groups around the world have adopted and/or extended the proposed methodology in their research. Spatiotemporal local binary patterns were later introduced and applied successfully to face analysis from video sequences for face and facial expression recognition, visual-speech recognition, lip reading, gender classification etc.
Figure 1: The basic LBP operator. It forms labels for the image pixels by thresholding the 3x3 neighborhood with the center value and considering the result as a binary number. The histogram of these 28 = 256 different labels can then be used as a descriptor.
Our more recent work focused on applying manifold learning techniques to face analysis (face recognition, age estimation, gender classification etc.) and on implementing face related applications in mobile phones. In the period 2008-2010, MVG participated in the Mobile Biometry (MOBIO) project funded by the European Commission and aiming at researching new mobile services secured by biometric authentication means. The LBP method developed in MVG played an important role in the project. We are currently continuing the biometry theme within a new FP7 consortium called TABULA RASA (starting in November 2010 and lasting 42 months). TABULA RASA looks at the vulnerabilities of existing biometric systems to spoofing attacks to a wide range of biometrics including face, voice, gait, fingerprints, retina, iris, vein, electro-physiological signals (EEG and ECG) etc.
Hadid A, Zhao G, Ahonen T & Pietikäinen M (2008) Face analysis using local binary patterns. In: Mirmehdi M, Xie X & Suri J (eds) Handbook of Texture Analysis, Imperial College Press, 347-373, (invited chapter).
Zhao G & Pietikäinen M (2007) Dynamic texture recognition using local binary patterns with an application to facial expressions. IEEE Transactions on Pattern Analysis and Machine Intelligence, 29(6):915-928.
Ahonen T, Hadid A & Pietikäinen M (2006) Face description with local binary patterns: Application to face recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 28(12):2037-2041.
Martinkauppi B, Hadid A & Pietikäinen M (2006) Color cue in facial image analysis. In: Lukac R & Plataniotis K (eds) Color Image Processing: Methods and Applications, CRC Press, 285-308, invited chapter.