We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Dynamic graph neural network-based computational paradigm for video summarization.
- Authors
Deepa, R.; Sree Sharmila, T.; Niruban, R.
- Abstract
In recent times, there has been a significant amount of interest in video summary technology. The reason behind video summarizing is to condense the information from the original video into a concise yet comprehensive summary, resulting in a brief and easy-to-understand version. In this study, looking at video summarization as an examination of graph issues and proposing a dynamic graph neural network to quantify the significance of individual videos and their overall contribution to the collection are covered. In prior graph network-based techniques, a single graph neural network (GNN) had often been used. The advantages of various graph filters or graph neural networks have not been fully exploited. The problem of oversmoothing still exists with traditional GNNs. To address issues like the one mentioned above, a spectral filter and an autoregressive moving average filter for the dynamic graph neural network have been proposed. In addition, a regularization of diversity to promote the model and provide a diverse summary has been recommended. Numerous tests have been run, and innovative and conventional graph models have been compared with the most advanced video summarizing techniques. The F-score for the two datasets got a big boost from 1.9 to 3.1% when the proposed method was used.
- Subjects
GRAPH neural networks; VIDEO summarization; MOVING average process; HYPERSPECTRAL imaging systems
- Publication
Multimedia Tools & Applications, 2024, Vol 83, Issue 17, p51227
- ISSN
1380-7501
- Publication type
Article
- DOI
10.1007/s11042-023-17412-4