A master's thesis at the College of Computer Science and Information Technology at University of Basrah explored masked face recognition using a Siamese neural network.

The thesis, presented by graduate student Ali Haitham Abdul Amir, aims to improve the performance of masked face recognition systems using a Siamese neural network by combining it with pre-trained models: Xception, MobileNet, EfficientNet, and ResNet-50, to extract deep features and compare images for identification.

The thesis included the use of data balancing to reduce bias and improve generalization on previously unseen data, enhancing model performance in real-world settings. The results demonstrated high effectiveness when combining the Siamese network with these models.