Face Verification System for Age-Invariance using Deep Convolution Neural Network(CNN)

There has been remarkable research progress made in the area of face recognition and it is been used in a wide range of applications from personal logins, security checks, attendance systems, etc., The progression of age reflects the facial features and this remains a great challenge while performing face recognition when matching is made with images of a large age gap. This paper proposes a robust face recognition technique invariant of the ages. The proposed system adapts the Deep Convolution Neural Networks (CNN) considering the matching of faces irrespective of the ages. Unlike other face recognition systems, the proposed system estimates the age and creates an Age-Invariant Face Representation using deep feature learning, where the features that are less affected by age-related changes like the texture Descriptors and the appearance-based features are extracted. This representation is used by the classification/matching step to perform recognition. The CACD data set is utilized to carry out the task of training and testing the algorithm. The system is verified to get robust results and when compared with the existing system provides an accuracy of 98.01% and the Rank-1 Accuracy of 88.21%.