PhD Dissertation Defense at the College of Information Technology, University of Babylon on an Integrated Intelligent Framework for Secure Video Processing

Duha Fadill Abbas
As part of its ongoing commitment to advancing scientific research and keeping pace with developments in artificial intelligence and cybersecurity, the College of Information Technology at the University of Babylon witnessed the defense of a PhD dissertation in the Department of Software by the student Farah Hussein Mohammed Jawad entitled “Disentangled Intelligence for Integrated Detection and Encryption in Secure Video Processing” under the supervision of Assistant Professor Dr. Mai Abdulmunem Saleh, in the presence of a number of faculty members, researchers, and postgraduate students, on Wednesday, 22/4/2026, at the College Conference Hall.

The dissertation addressed a contemporary challenge arising from the increasing demand for video processing on resource-constrained edge devices, where a significant trade-off exists between the high computational requirements of deep learning–based object detection and the necessity of robust data encryption. Existing systems often suffer from latency and hardware bottlenecks when attempting to perform both tasks simultaneously, particularly due to the monolithic architectures of conventional object detection frameworks.

The study aimed to design a unified framework based on a disentangled intelligence model to enhance object detection efficiency while reducing resource consumption and ensuring a high level of data security. A lightweight model was developed through the optimization of convolutional neural network architectures, introducing an improved regression head to simplify spatial localization and enhance detection accuracy.

In parallel, the research proposed an enhanced stream cipher algorithm termed EChaCha, in which the number of encryption rounds was reduced to twelve to improve computational efficiency. This reduction was compensated by integrating SHA-256 to dynamically modify rotation constants, in addition to incorporating a chaotic logistic map for nonlinear block permutation, thereby significantly strengthening cryptographic robustness and resistance to attacks.

This dissertation presents an advanced and specialized architecture that bridges the gap between efficiency and security in video processing, reinforcing the research direction of the College of Information Technology toward developing innovative solutions capable of addressing modern technological challenges and enhancing data security in advanced digital environments.