3D Face Recognition Using PCA Based Deep CNN For Left And Right Half Face
Currently computer-based face recognition is a sophisticated and robust system that is
basically being used for several cases of access control. In case of face detection or
authentication which has been conducted primarily using full form of facial image data. Face
recognition is a challenging problem based on facial features which are aimed at establishing a
person’s identity because of dissimilar facial manifold nature. The application in face
recognition has wide range based on labeling strategies and the class label belongs to the
unknown class which has been assigned to the test image. In this paper described the intrinsic
facial symmetry features for face recognition based on Left Half Face (LHF) and Right Half
Face (RHF) using Principal Component Analysis (PCA) algorithm to learn mapping in 3D face
images using deep Convolutional Neural Network (dCNN). The experimental results have shown
that dCNN based PCA is used for the recognition and classification based on MORPH dataset.
The overall performance has assessment in terms of number of training faces on the original
image, left and right faces data. Moreover, the PCA-dCNN recognition performance is better
than the vectors of feature projection obtained through the PCA based approach.