A Master thesis was discussed in Department of Electrical Engineering / College of Engineering at University of Mosul entitled "Design an Object Detection and Recognition System Based on Deep Learning" submitted by (Ayad Saadi Ahmed Almamary), Supervised By Assist. Prof. Dr. Mohammed Obaid Mustafa. On Thursday, Nov. 24, 2022.

The CNN (Convolutional Neural Networks) algorithm used with YOLO (You Only Look Once) and Faster RCNN algorithms, which are based on the (CNN) algorithm to identify objects.

A set of practical data with a number of classes and different capturing conditions, the training was done using the fifth version of the YOLO algorithm, which is the latest version, which was released in 2020. It was built using the Python language, also we build a model using Faster RCNN, and we made a comparison between the two models, then the models were tested in images that the models had not previously seen, as they gave high efficiency to discover objects and determine their types, The second algorithm outperformed in terms of accuracy, but was slower in detecting compared to the first then YOLO was tested on video with the same lighting conditions for the training data and inside a moving car where it came distinguished and close to the data taken with better imaging conditions.

Also Read