H. Sellahewa & S.A. Jassim, “Image-quality-based adaptive face recognition”
H. Sellahewa & S.A. Jassim, “Image Quality-based Adaptive Face Recognition”, in IEEE Trans. on Instrumentation and Measurement, vol. 59 (4), pp. 805–813, April 2010.
The accuracy of automatic face recognition systems is greatly affected by varying lighting conditions between enrolment and recognition stages. This problem is addressed by applying illumination normalisation techniques on face images and/or by using illumination invariant face feature representations for classification. The paper presents a context-aware adaptive face recognition scheme, where the selection of normalisation technique as well as the face feature representation is based on the illumination quality of a given face image.
The proposed algorithm extends Dr Sellahewa and Professor Jassim’s previous work on wavelet-based multi-stream face recognition and is a part of the Department of Applied Computing’s ongoing research into developing context-aware adaptive biometric systems for mobile devices.
An abstract of the paper is available from the IEEE website (external link).
Harin Sellahewa has a PhD from the University of Buckingham and is Research Lecturer in the Department of Applied Computing. Sabah Jassim is Professor of Mathematics and Head of the Department of Applied Computing at Buckingham.