Fast self-quotient image method for lighting normalization based on modified Gaussian filter kernel

ABSTRACT

The Self-Quotient Image (SQI) Method [Wang H, Li SZ, Wang Y, et al. Self quotient image for face recognition. International Conference on Image Processing (ICIP’04); 2004;Vol. 2. p. 1397–1400; Wang H, Li SZ, Wang Y. Generalized quotient image. IEEE CVPR; 2004; Vol. 2. p. 498–505] is a simple method for lighting normalization based on the Quotient Image method [Shashua A, Riklin-Raviv T. The quotient image: class-based re-rendering and recognition with varying illuminations. T Pattern Anal Mach Intel. 2001;23(2):129–139; Riklin-Raviv T, Shashua A. The quotient image: class based recognition and synthesis under varying illumination. Proceedings of the 1999 Conference on Computer Vision and Pattern Recognition; 1999; Fort Collins (CO). p. 566–571]. The main advantage of the SQI is the use of only one image for lighting normalization. Nevertheless, the SQI still has few disadvantages which make hard to use it in some face recognition systems. In this paper, we introduce the modified version of the SQI method based on globally modified Gaussian filter kernel. In this modification, we tried to solve the disadvantages of the original SQI method, simplify the computational process, and increase the quality of illumination normalization. We have investigated two modification of the original SQI method and shown how they normalize different shadow regions.

KEYWORDS:

• Self-quotient image
• SQI
• quotient image
• Gaussian filter
• face recognition
• lighting normalization
• illumination normalization

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Vitalius Parubochyi graduated from the Ivan Franko State University of Lviv with honours in 2015. Has been working at the Ivan Franko National University of Lviv since 2015. In 2015, he successfully completed the Joint Certificate Program on Advanced Computational Mathematics, Ivan Franko National University Lviv, Ukraine and Julius Maximilian University Würzburg, Germany. From 2015 to 2018, he did his postgraduate study at the Ivan Franko National University of Lviv. His research interests are High-performance computing, Parallel computing, Computing using MPI, CUDA, OpenCL and hybrid architectures, Data science, Machine learning, Image processing and recognition, Neural Networks, and Real-Time Recognition Systems for low-performance systems.

Roman Shuwar graduated from the Ivan Franko State University of Lviv with honours in 1982. He has been working at the Ivan Franko National University of Lviv since 1982. During 1985-1988, he did his postgraduate study. In 1989, he was Candidate of Physical and Mathematical Sciences; from 1988 to 1994, he was Assistant Professor; from 1994 to till now, he is Associate Professor; and from 2016 to till now, he is also Acting Chairperson of the System Design Department of the Faculty of Electronics and Computer Technologies. His research interests are High-performance computing systems, Algorithms of parallel computing, Computing using MPI, CUDA, Grid and hybrid architectures, Data science, Machine learning, Intelligent image processing systems, and Optimization methods.

ORCID

Vitalius Parubochyi http://orcid.org/0000-0002-9847-6968

Image notes

The image dataset used for figures 1 to 5 can be accessed through the following links:

https://computervisiononline.com/dataset/1105138686

http://vision.ucsd.edu/∼iskwak/ExtYaleDatabase/ExtYaleB.html

http://vision.ucsd.edu/∼leekc/ExtYaleDatabase/Yale%20Face%20Database.htm

Images in figures 6 and 7 have been reproduced with permission from the authors of Ref. [16]