Illumination normalization using independent component analysis and filtering

ABSTRACT

In this work, we separate the illumination and reflectance components of a single input image which is non-uniformly illuminated. Considering the input image and its blurred version as two different combinations of illumination and reflectance components, we use the conventional independent component analysis (ICA) to separate these two components. The separated reflectance component, which is an illumination normalized version of the input image, can then be used as an effective pre-processed (illumination normalized) image for different computer vision tasks e.g. face recognition. To this end, we present simulation results to show that our proposed pre-processing method called illumination normalization using ICA increases the accuracy rate of several baseline face recognition systems (FRSs). The proposed method showed improved performance of baseline FRSs when using the Extended Yale-B databases.

KEYWORDS:

• Illumination
• reflectance
• illumination normalization
• reflectance estimation

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Fawad Ahmad is a Ph.D. student in Department of Electronic Engineering, City University of Hong Kong, Kowloon, Hong Kong. He received the B.S. degree in Electrical Engineering from the National University of Computer and Emerging Sciences, Peshawar, Pakistan, in 2013 and M.S. degree Information, Communication and Electronics Engineering from the Catholic University of Korea, Bucheon, Korea.

Asif Khan is assistant professor with Ghulam Ishaq Khan (GIK) Institute of Engineering Sciences and Technology, Swabi, Pakistan. He received Ph.D. degree in Interactive and Cognitive Environments jointly awarded by the three universities: Alpen-Adria-Universität Klagenfurt, Klagenfurt, Austria, Queen Mary University of London, London, UK, and Università degli Studi di Genova, Genova, Italy in 2015. His main research interests include image processing and multi-robot systems.

Ihtehsam Ul Islam received Ph.D. degree in Computer and Control Engineering from Politecnico Di Torino, Torino, Italy in 2015. His current research interests include computer vision, machine learning, and image processing.

Muhammad Uzair received Ph.D. degree in computer engineering from the University of Western Australia (UWA), Crawley, WA, Australia, in 2016. His current research interests include computer vision, machine learning, domain adaptation, image set modelling, and hyperspectral image analysis.

Habib Ullah did Ph.D. in 2015 from the University of Trento, Trento, Italy and the M.Sc. degree in Computer Engineering in 2009 from Hanyang University, Seoul, South Korea. In 2015, he received a prestigious competitive Postdoctoral fellowship from Ecole De Technologie Superieure, Montreal, Canada. His research interests include computer vision, crowd motion analysis, video surveillance, and machine learning.

ORCID

Asif Khan http://orcid.org/0000-0002-9840-5289

Notes

All images obtained from The Extended Yale Face Database B, which can be accessed through the following link: http://academictorrents.com/details/06e479f338b56fa5948c40287b66f68236a14612