CNN based Smart Surveillance System: A Smart IoT Application Post Covid-19 Era

摘要

There are many applications available in the face detection algorithm but very limited applications are identified for further processing. When it comes to identifying faces in the crowd and that too in all-weather conditions then it's too difficult a task to be conducted. Considering this challenge, most of the surveillance systems are not automated. In the sense the CCTV deployed are used only for bufferingpurposes. Very rarely an event is brought to the notice and later CCTV footage is used as a tool for legal issues. There are a lot of concerns about deploying CCTV's in Public places as well. Selecting this system, our work have made sure that the CCTV's will be used to process the video being taken, and whenever an event triggers, notice(s) has to be provided for officials. Thus, the proposed system is for automatic detection and recognition of human faces for finding criminals/suspects/missing persons for surveillance. The intended method first detects a face in the video using a face detection algorithm and then it will search whether the face is present in the data centre. The method provides the ability to detect, extract features, and recognize a face from inputs taken by camera or video automatically. Recognizing faces under different natural conditions can be done by training the system on a limited number of facial images. Also, this system is tested for Covid-19 Post situations, wherein wearing masks in public places is compulsory thus face recognition in these aspects is also tested, and encouraging results have been achieved.