Fourier Spectrum Features for Face Recognition
Dr..X.Ascar Davix, C. John Moses, Suresh Kumar Pittala, D. Eswara Chaitanya · International Journal of Innovative Technology and Exploring Engineering · 2020
The objective is to introduce a novel approach which deals with the challenges: uneven illumination and partial occlusion. This method performs face recognition by extracting the magnitude spectra features. At each point on the face, largest matching areas were found. Thus robustness is achieved using Fourier magnitude spectra feature extraction and largest matching area comparison. This method performs competitively with corrupted images and other unsupervised methods. The proposed approach is experimented on Yale B and AR datasets.