Project Abstract

Video Summarization addresses the expanding abundance of video data by presenting aninnovative method for automatically analyzing long videos from numerous perspectives andproducing concise, useful summaries. The system uses cutting-edge deep learning models,particularly Long Short-Term Memory (LSTM) networks, to implement a novel progressivesummarization technique based on self-supervised learning of multimodal inputs. This iterativetechnique enhances summarization quality while minimizing redundancy and reliance on labor-intensive annotations, leading to greater scalability and efficiency. Extensive trials on benchmarkdatasets show the system’s excellent performance, which is aided by the progressivesummarizing technique. Users may instantly receive customized summaries through an intuitiveinterface, changing the way they consume video material. By increasing accessibility andengagement, the technology streamlines information extraction from video,revolutionizing multimedia analysis and benefiting a wide range of applications.

Keywords: Video Summarization, Automated Video Summarization, Deep Video Summarization

 Conference Details

Session: Presentation Stream 17 at Presentation Slot 4

Location: GH011 at Wednesday 8th 09:00 – 12:30

Markers: Arnold Beckmann, Alex Warren (GTA)

Course: MSc Data Science, Masters PG

Future Plans: I’m looking for work