College of Education for Pure Sciences, University of Thi-Qar discussed a master dissertation on the adaptive system for controlling traffic lights based on artificial intelligence by the postgraduate student, Ms. Hind Ali Nasser Hussein.
The dissertation aimed at implementing a system based on image processing that uses technology to detect objects in the image based on convolutional neural networks (CNN), determine traffic intensity based on the number of objects and control traffic signals according to traffic density on each road.
The dissertation highlighted that the proposed system was implemented using the MATLAB R2022 language and to evaluate performance, classification accuracy was used as a measure of evaluation on a lot of traffic.
The dissertation concluded that using the (ResNet50) model led to a significant improvement in performance and that the proposed system using (ResNet50) achieves an average accuracy of up to 99.47% on (MTID) data sets in consequence.
