Shearlet based CRLBP Texture Features for Recognition of Facial Expression

Nagaraja S.

Abstract: In this research article, we introduced a method for capturing facial expressions by extracting texture features using the Completed Robust Local Binary Pattern (CRLBP) in the shearlet domain. The shearlet transform shows better directional capabilities and a stronger ability to represent edges and other singularities along curves compared to conventional multiscale transforms like the wavelet transform. Therefore, we convert the original face images into the frequency domain by applying the shearlet transform at a particular scale and orientation. Texture features that are unaffected by noise and illumination are obtained from an approximate sub-band through the use of completed robust local binary pattern. These features are then combined to create a feature vector representing facial expression. The suggested approach is assessed by conducting facial expression recognition on a benchmark database like JAFFE. The process of identifying facial expressions involves the use of a chi-square distance measure and a nearest neighbour classifier. The experimental results demonstrate that our method performs better than other well-known LBP-based methods.

Keywords: CRLBP, Shearlet transform, Facial Expression, Texture features