Originally Posted by DaFace:
I hadn't thought about that one. You'd theoretically need an entire ring of sensors plus one in each point. Again, you could probably solve for that, but the more complex you make it, the more likely it'll be screwy at times.
Words like "Optical Tracking" sound like they're playing around with computer vision, AKA convolutional neural networks and drawing bounding boxes. Using that kind of technique you wouldn't need sensors you'd use a model that's been trained on tens of thousands of hours of video and you measure the probability of false alarm and probability of detection. Get the PD up to 99% and the PFA down and then you can have some level of confidence that you're outperforming the human.
[Reply]