A Novel Approach to Prevent Unauthorized Access and Enhancement in Security of Database Template
Face Recognition is considered to be the most suitable technique for real time application. This technique is
commonly used for the security purposes for authentication in computerized fields. Previously various algorithms and
techniques were used for the purpose of security and authentication but they were having several pitfalls like time
consumption, pose and illumination problem along with age differences. One more important point to note was that the
image in the database was also not encrypted so chances of breach in security would have been easily created. Keeping these
things in mind along with the available literature review a hybrid technique for face recognition with encryption security is
proposed in this work. In the proposed method, face recognition is done by combining two most commonly used techniques
and making it hybrid in nature which would combine their advantages and reduce the false matching rate along with fast key
generation time. Moreover for security purpose encryption and decryption of the database is also done. Local Binary Pattern
and Principal Component Analysis techniques were combined in our hybrid proposed technique. PCA reduce dimension of
image and LBP describe the local texture of image data. So proposed method will increase the recognition rate and decreased
false match rate. So It is more suitable for real time application.
Keywords- Principal Component Analysis (PCA), Local Binary Pattern (LBP), Eigenfaces, feature extraction, face
recognition, Cryptography, Recognition Rate (RR), False Match Rate (FMR).