TAILORED VIDEO SUMMARIZATION: CATERING TO PATIENT AND IMPATIENT USERS

People across various fields rely on automated video summarization tools to manage extensive video content efficiently. This research focuses on developing a dynamic, user-centered approach to video summarization, accommodating both patient and impatient user needs. The system aims to handle lengthy videos by identifying and cataloging all objects within them. It follows a three-step process: Object-of- Interest selection, object detection/localization, and video summarization. For patient viewers, it offers comprehensive scene identification and storage. For impatient users, it provides concise summaries quickly. By adapting itself to individual preferences, this research will make videos more accessible and useful by providing personalized video summaries which will help avoid information overload in various spheres such as security, entertainment, or personal documentation. This research used deep learning models like YOLOv8, ResNeXt as well as LSTM to implement this user- centric approach to video summarization.