Security professionals have been paying close attention to face recognition (FR), particularly when it comes to using closed-circuit television (CCTV) cameras for security surveillance. Even though the science of computer vision has made great strides, sophisticated face recognition algorithms have only performed well under controlled circumstances. When faced with real-world situations like illumination, motion blur, camera resolution, etc., they drastically decrease. The design, implementation, and empirical comparisons of machine learning open libraries in creating attendance taking (AT) Support systems employing interior security cameras, or ATSS, are demonstrated in these papers. Our design enables flexible system scaling and can be used for general school attendance with CCTV. The measurement results demonstrate that the accuracy is appropriate for a wide range of settings