M2VTS

Multi-modal Verification for Teleservices and Security Applications web page at Aristotle University of Thessaloniki

Project Objectives

M2VTS project addresses the issue of secured access to local and centralised services in a multimedia environment. The major problem in user authentification is to achieve on the one hand toll performance: false acceptance rate as low as possible (minimise access to impostors), and false rejection rate as low as possible (a registered user should access to his system in any case), and on the other hand stand the wide range of conditions of use of such systems as well as to provide ergonomically viable solutions. The use of multiple modalities will help to overpass potential technical limitations of individual modalities as well as take benefit of the emerging multimedia environment : workstations, network computers, smart phones are more and more equiped with audio and video capabilities. The research is driven by the application needs and user requirements. Therefore, work has essentially been driven by three main goals:

• Develop platforms for evaluation, implementation and fast prototyping of technology. Submit these platforms to user tests in real situations, in order to measure the adequation between user requirements and current maturity of the technology.
• Develop algorithmic solutions for user authentification in a multi modal context. Implement these solutions on the software platforms for fast prototyping. Refinements of the algorithms based on the results of the user tests in real situations.
• Develop prototypical applications for end users.

Research topics addressed by our team

• face detection, facial feature extraction and tracking ,
• face normalization ,
• frontal face verification based on morphological dynamic link architecture ,
• frontal face verification based on morphological signal decomposition – dynamic link architecture ,
• frontal face verification using morphological pyramids ,
• fusion.

Papers

• K. Sobottka, I. Pitas. Localization of facial regions and features in color images. Fourth Open German-Russian Workshop on Pattern Recognition and Image Understanding. Valday, The Russian Federation, March 1996.
• K. Sobottka and I. Pitas. Looking for faces and facial features in color images. Full paper published in Pattern Recognition and Image Analysis: Advances in Mathematical Theory and Applications. Russian Academy of Sciences, 1996.
• M. Acheroy, C. Beumier, J. Biguen, G. Chollet, B. Duc, S. Fischer, D. Genoud, P. Lockwood, G. Maitre, S. Pigeon, I. Pitas, K. Sobottka and L. Vandendorpe. Multi-modal person verification tools using speech and images. European Conference on Multimedia Applications, Services and Techniques. Louvain-La-Neuve, Belgium, May 1996.
• K. Sobottka and I. Pitas. Extraction of facial regions and features using color and shape information. International Conference on Pattern Recognition (ICPR 96), Austria, September 1996.
• K. Sobottka and I. Pitas. Segmentation and tracking of faces in color images. Second Int. Conf. on Automatic Face- and Gesture-Recognition 96, Killington, Vermont, USA, October 14-16, 1996.
• K. Sobottka and I. Pitas. A novel method for automatic face segmentation, facial feature extraction and tracking ( pp. 1-18 , pp. 19-22 , pp. 23-34 ), Signal Processing: Image Communication vol. 12, pp. 263–281, 1998.
• C. Kotropoulos and I. Pitas. Rule-based face detection in frontal views. 1997 IEEE Int. Conf. on Acoustics, Speech and Signal Processing, Munich, Germany, April 20-24, 1997.
• C. Kotropoulos, I. Pitas, S. Fischer, and B. Duc. Face authentication using Morphological Dynamic Link Architecture. First Int. Conf. on Audio- and Video- based Biometric Person Authentication (AVBPA 97), Crans-Montana, Switzerland, March 1997.
• K. Sobottka and I. Pitas. A fully automatic approach to facial feature detection and tracking. First Int. Conf. on Audio- and Video- based Biometric Person Authentication (AVBPA 97), Crans-Montana, Switzerland, March 1997.
• C. Kotropoulos and I. Pitas. Face verification based on morphological dynamic link architecture. 1997 IEEE/EURASIP Workshop on Nonlinear Signal and Image Processing (NSIP 97), Mackinac Island, Michigan, USA, September 8-10, 1997.
• C. Kotropoulos and I. Pitas. Face authentication based on morphological grid matching. 1997 IEEE Int. Conference on Image Processing (ICIP 97), Santa-Barbara, California, USA, October 26-28, 1997
• A. Tefas, C. Kotropoulos and I. Pitas. Face verification based on morphological shape decomposition. 3rd Int. Workshop on Face and Gesture Recognition, Nara, Japan, April 14-16, 1998
• C. Kotropoulos, A. Tefas and I. Pitas. Face authentication using variants of elastic graph matching that incorporate local discriminant coefficients. 1998 IEEE Int. Conference on Acoustics, Speech and Signal Processing (ICASSP 98), Seattle, Washington, U.S.A., May 12-15, 1998
• A. Tefas, C. Kotropoulos and I. Pitas. Variants of Dynamic Link Architecture based on mathematical morphology for frontal face authentication. 1998 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 98), Santa Barbara, California, U.S.A., June 23-25, 1998
• V. Chatzis, A.G. Bors and I. Pitas. Decision Level Fusion by Clustering Algorithms for Person Authentication. 1998 European Signal Processing Conference, Rhodes, Greece, September 8-11, 1998
• C. Kotropoulos and I. Pitas. Linear Projection algorithms and morphological dynamic link architecture for frontal face authentication. 1998 European Signal Processing Conference, Rhodes, Greece, September 8-11, 1998
• C. Kotropoulos, A. Tefas and I. Pitas. Frontal face authentication using variants of dynamic link matching based on mathematical morphology. 1998 IEEE Int. Conference on Image Processing, Chicago, U.S.A., October 4-7, 1998