If it’s mostly about motion detection and false positives, it’s hard to avoid the object detection stuff if you aren’t doing that already. I stopped using Blue Iris before it started integrating object detection and I can’t speak to it at all. I liked the setup I had with ZoneMinder, Zoneminder Event Server, and the associated machine learning hooks, but the developer behind the event server and ML stuff left and it isn’t particularly well integrated. I like how quickly and comprehensively Viseron (github.com/roflcoopter/viseron) has progressed. Very sleek software, engaged developer, uses YOLOv7, and even supports an openvino backend. Well worth a look.