Extraction of Texture Features from Shearlet Face using RLBP for Recognition
Srinath Nagaraja · International Journal of Science and Research (IJSR) · 2024
In this paper, we introduced a method for representing faces using a combination of Shearlet Transform and Robust Local Binary Pattern (RLBP).The shearlet transform exhibit improved multiscale directional capability, and a greater ability to localize distributed discontinuities such as edges along curves as compared to traditional multiscale transform such as wavelet transform and curvelet transform. Thus, we convert the initial face images into the frequency domain by utilizing the shearlet transform. Noise-robust binary patterns are obtained from the approximate sub-band through the use of the RLBP technique in order to symbolize the face. These patterns create the feature vector. The effectiveness of the suggested approach for representing faces is assessed by performing face recognition on a benchmark database like ORL. The technique Support Vector Machines (SVM) with a polynomial kernel function is employed to perform classification. The experimental findings indicate that our method performs better when compared to approaches based on LBP.