Computer hardware and software systems for computer and machine learning, specifically intended for autonomous learning, or training of the machine, to recognize specific visual objects within a scene