Human Emotions Detection using Kernel Nonlinear Collaborative Discriminant Regression Classifier : Human Emotions Detection using KNCDRC

This paper has proposed a non-linear subspace classifier that uses a linear collaborative discriminant regression classification method on a kernel version called Kernel Nonlinear Collaborative Discriminant Regression Classification method (KNCDRC). The kernel method is generally considered helpful in determining non-linear Structure. Unlike traditional classification methods, the proposed model use the kernel function to plot the test points from the low dimensional kernel space into the high dimensional kernel space. The proposed method is to detect human emotion through facial feature recognition. To improve recognition rate and accuracy, a future subspace has been created on the dataset. It enhances the connection between sample points and the dimensions of the sample points. The experiments show that KNDCRC achieves a better recognition rate than existing state-of-art recognition systems on large-scale standard datasets like Multi-PIE and AffectNet.