Face has recently attracted more attention in a number of fields because to its security features and its simplicity of use. Face-based biometric technologies are widely used in various person identification applications. This is due to the fact that human faces are the most easily identifiable features from day-to-day living and also store the most information. However, despite ongoing attempts to spoof faces, facial recognition systems continue to be vulnerable to attack. The act of damaging or attacking a face recognition system by gaining unauthorized access to the system and exploiting security holes is known as spoofing. This may be accomplished by getting into the system without the agreement of an authorized user. The act of damaging or attacking a face recognition system by gaining unauthorized access to the system and exploiting security holes is known as spoofing. This may be accomplished by getting into the system without the agreement of an authorized user. Attacks that involve faking one's visage provide a persistent risk to face-recognition systems. Our goal is to create a system that will put an end to face spoofing, despite the fact that academics have created a variety of face spoofing detection approaches that have demonstrated to be highly effective. Anyone is capable of fooling a facial recognition system by uploading fake photos or videos of themselves, or by employing some other decoy to stand in for an authorized user's face. The proposed work, which is based on an algorithm for a deep neutral network, suggests the real spoofing prevention method by assessing the liveness of the face. Additionally, it offers defense against spoofing attacks such as image masks, replay assaults, print photo assaults, and mobile photo assaults. The technique has a success rate of 99%, particularly when applying the convolutional neural network (CNN) algorithm. In facial recognition systems, the primary goal is to accurately differentiate between real and fake faces using the CNN approach. This is done to improve accuracy.