Anybody has experimented with or has any recommendations on how to deal with this? I had the opportunity to test Dahua and Hikvision cameras. They both have really nice "smart features" as they are able to detect when: "Something" is crossing a line. "Something" is entering or leaving an area. "Something" is moving through the picture. However, "something" could be a person, tree leaves, a cat wondering around or even a shadow from a tree. While I can see these features being bulletproof indoors (where the only thing moving should be a person), they become almost useless outdoors (for me at least) due to trees, wind, animals and even the occasional spider crawling in front of the camera. So I was wondering: Isn't there already a solution that can detect human body and alert me only - and if only - a person is present? If not, how are you remediating false positives? With today's technology the perfect solution is so "easy" to build that I almost can't believe a it doesn't already exist. If I can't find anything I might build it, the principle is simple (using AI): Tap in one of the camera's stream or NVR (low res, low frame rate ) Continuously run each frame through a pre-trained model Take action if it detects a person, an animal or both in at least one frame. Actions could be: Communicate with NVR so it can bookmark the video or use its built in alert system Send an email or text message with the video frame. Depending on time of the night, turn lights on (through integration with smart home) and... wake me up so I can take a shoot at the bastard The goal is simple: Be aware of people wondering around my house while filtering out all false positives.