Suppose you want to detect objects and also track objects and people. Is it better to train a model using a single fisheye camera or using multiple cameras that mimic the view of the fisheye camera?
Also, what can be done to remove objects that are washed out? Like for very small objects, how do you make them more visible? Would multicamera tracking be better in this scenario?