In a time of increasing dependence on digital identity, facial biometric systems have become important for access control and authentication. However, their vulnerability to advanced presentation attacks is especially 3D masks and media -based (photos, videos and replay’s like gif) reflects serious security threats. Due to liveness detection will show some impact on the major threats which are unseen attacks. This literature survey systematically examines the development of real -time detection techniques developed to combat these challenges. We analyze deep learning methods and hybrid models such as classic texture-based methods, 3D depth and thermal sensing, CNN and VITs, customized distribution on edge units. The paper data set discusses availability, real -time performance, generality of the domain and hardware efficiency. Research intervals have been highlighted in adapting cross -film, unfavorable strength and distribution of low resources. This review serves as a basis for researchers and doctors, who aim to develop flexible, real-time anti-spoofing systems for safe biometric authentication. Apart from this some of review highlights the importance of standardized evaluation protocols, privacy-preserving deployment and fairness across demographics groups.