Face Recognition Based on Local Image Descriptor and Non-linear Features Extraction
Mina Hojjati
Mina Hojjati, Department of Electrical and Computer Engineering, Science and Research Branch, Islamic Azad University, Qazvin, Iran.
Manuscript received on October 04, 2013. | Revised Manuscript Received on October 17, 2013. | Manuscript published on October 20, 2013. | PP: 27-30 | Volume-1, Issue-11, October 2013. | Retrieval Number: K03271011113/2013©BEIESP
Open Access | Ethics and Policies | Cite
© The Authors. Published By: Blue Eyes Intelligence Engineering & Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: In this paper, we introduce EABF (Extraction Analysis of Bsif Features) new method to face recognition based on extraction and analysis of binary sif features (BSIF). In our proposed algorithm, FABF eliminates some objections that led to many problems in the previous algorithms, such as a large query space and different quality and the size of images due to different time conditions for imaging and it removes the disadvantages of ELPDA (Nearby Local Discriminating Analysis) methods as a between-class-Scatter by using the Scatter matrix. This matrix introduces and updates the nearest neighbors to the outer class (K) through the samples. In addition, one of the advantages of the EABF is the high-speed face recognition by reducing the size of feature matrix and using NLPCA (Non-Linear Locality Preserving Analysis). Finally, the experiments results on the FERET data base indicate the impact of proposed method on the face recognition.
Keywords: Face Recognition, Non-linear Features, Linear Features, Local Image Descriptor.