Significant progress in automated face recognition systems during the last ten years has changed the landscape of security applications. However, obstacles still exist, impeding broad adoption, especially in sectors where security is a concern. The impact of non-permanent facial cosmetics on automated face recognition is an important but little-examined topic in this study. To examine the effect of makeup on recognition accuracy, we carefully select two databases: pre- and post-makeup application facial photos. We highlight the possible vulnerability it introduces and the pressing need for better understanding by revealing its large effect through extensive experimentation. Moreover, we expand our study to include image impersonation detection, a crucial issue for face recognition systems. We create two models: one for spotting makeup and another for telling authentic photos apart from fakes. These models show how to proactively strength biometric systems against threats of impersonation. By creating strong algorithmic solutions that can detect image impersonation and navigate hurdles caused by cosmetics, our study contributes to the integrity and dependability of face recognition systems in security scenarios.