Detection, identification and intelligence are the blistering topics and requirements nowadays in modern era. As artificial intelligence is going to seizure the whole world soon, most of the explorations are going towards it. Such systems have applications in numerous fields of life such as security surveillance, authentication etc. In human face detection and identification systems, there are many hurdles that could lead the system to fail, such as wearing cap and glasses hide face and eyes respectively from investigation camera. Therefore, glass recognition and cap from investigation video has converted a challenging issue in computerization of safety systems. This research is principally focused on disturbances in the process of detecting faces. Due to these challenges an accurate and fast system that can detect in various situations in real time is needed. We have proposed a detection algorithm that works well in occlusion. This algorithm first detect faces using regions of interest (ROIs), then detects eyes pair and differentiates eyes pair from mask or glasses and exploit the blob analysis option in case of mask or glasses under surveillance. Then detects the head, if the head is detected it is highlighted else blob analysis is performed. The results shows that our proposed methodology performs better in circumstances of occlusion, and increases the efficiency up to 98% on haar features detection based face detection, while 93% on occlusion.