A Local Descriptor for Age Invariant Face Recognition under Uncontrolled Environment
Rajesh Kumar Tripathi, Anand Singh Jalal · 2021 2nd International Conference on Secure Cyber Computing and Communications (ICSCCC) · 2021
Face recognition has major applications such as missing people or children identification or criminal identification. These applications become more challenging while having age variation. To handle the above applications, an age invariant face recognition system is required. This paper proposes a robust descriptor which works on local region of face parts and extract the discriminative features to identify the people. In the first step, face detection, alignment and cropping is done, secondly, proposed discriminative descriptor is utilized over the local regions of eye, nose and mouth parts. In the third step, a machine learning method -Support Vector machine is utilized for the classification. Finally, Rank-1 recognition accuracy is computed with 5-fold method. To measure the effectiveness of the proposed descriptor, a standard and challenging dataset has been utilized. The effectiveness of the descriptor has been proved in the result section and proposed descriptor outperformed to the existing descriptor based approaches.