Ph.D. Dissertation at the College of Information Technology Discusses "Static Video Summarization Based on Trajectories Behavior"

Duhaa Fadill Abbas
The Software Department at the College of Information Technology – University of Babylon held a Ph.D. dissertation defense entitled "Static Video Summarization Based on Trajectories Behavior", presented by doctoral candidate Alya Talib Raheem Najm and supervised by Prof. Dr. Tawfiq Abdulkhaliq Abbas. The defense took place at 9:00 a.m. on Tuesday, July 1, 2025, in the conference hall of the college.

The dissertation highlights that in recent years, there has been a rapid global increase in the volume of digital video content and storage capacity due to the widespread use of visual data. This expansion necessitates the development of efficient and rapid video processing techniques. Among such techniques is content-based video retrieval, in which video summarization serves as an effective solution for managing and browsing massive volumes of video content. It also provides a powerful means of reducing storage requirements and accelerating search and retrieval processes. Consequently, there is an urgent need for techniques capable of presenting video content in a concise form without compromising essential information. Video summarization has emerged as a key technique in this domain and has become a highly active research area in recent years.

Video summarization refers to the process of generating a condensed version of a video that retains the most informative and representative content from the original footage. The goal is to accurately convey the essence of the video and provide meaningful insights to the user. Such summarization processes often require temporal segmentation of the video into its basic units—sequential frames. Accurate temporal segmentation enhances video analysis and leads to improved outcomes. Various techniques have been proposed for this purpose, ranging from classical computer vision methods to deep learning-based approaches.

The dissertation also explores methods for both static and dynamic video summarization using deep learning techniques and other algorithms that rely solely on the motion trajectory characteristics of objects within the video.