College of Science, Mustansiriyah University discussed a Ph.D. thesis on designing of a system for detecting and recognizing humans in video scenes using the (RCNN) algorithm by the postgraduate student, Mr. Majed Kamel.
The thesis aimed at adopting the (RCNN) algorithm, benefitinf from it and employing it in building an effective system for detecting humans in video scenes at multiple different distances starting from (10) meters up to (120) meters in local environments inside Iraq and in varying circumstances.
The thesis reviewed improving the (RCNN), (Fast RCNN) and (Faster RCNN) techniques in the field of human detection and designing the new system with several models, as well as testing these techniques with image samples of a group of people positioned at various distances in order to enable the models to detect about people over a wide range of distances.
The thesis concluded that the model based on (Faster RCNN) technology is the most effective in detecting humans, as its accuracy rate was (100%) for distances ranging between (10) meters and (50) meters, while the percentage decreased slightly to reach (88%) in detecting humans at a distance of (120) meters
