This paper presents an illumination invariant face recognition technique that uses a combination of local edge gradient information from two different neighboring pixel configurations to represent face images. The proposed Local Boosted Features (LBF) is an oriented local descriptor that is able to encode various patterns of face images under different lighting conditions. It employs the local edge response values in different directions and multi-region histograms from each neighborhood size. Then concatenate these histograms to get one long LBF-feature vector for each image. Finally, we use a library for support vector machines (LIBSVM) classifier to define the similarity between a test feature vector and all other candidate feature vectors. The performance evaluation of the proposed LBF algorithm is conducted on several publicly available face databases and observed improvements in the recognition accuracy.
Almabrok Essa, Vijayan K. Asari, "Local Boosted Features for illumination Invariant Face Recognition" in Proc. IS&T Int’l. Symp. on Electronic Imaging: Imaging and Multimedia Analytics in a Web and Mobile World, 2017, pp 70 - 73, https://doi.org/10.2352/ISSN.2470-1173.2017.10.IMAWM-170