Deep Learning Models for Biometric Recognition based on Face, Finger vein, Fingerprint, and Iris: A Survey

Recently, individual biometric characteristics gained a lot of attention and are the heart of building multiple kinds of security and authenticity systems, such as surveillance, forensic, fraudulent disclosing, and identity-based access control. The vast types of biometrics traits make the procedure of selecting the suitable one a crucial issue, which mainly depends on the type of application, the availability of samples, the degree of intricacy, and the accepted value of possibility. The concept of machine learning algorithms has gained a big interest in the last manner, especially the evolved version of it named as deep learning neural networks. Machine learning has been utilized and implemented in a lot of biometric systems due to its powerful properties and capabilities which can provide the desired goal from the system with great performance. This work serves to introduce an extensive survey of more than 190 promising works from the past seven years that describe multiple kinds of biometric-based deep learning systems based on four popular and most utilized traits of great characteristics, including face, fingerprint, iris, and finger vein. A brief review of both biometrics' kinds, and deep learning neural networks is also presented in this article.