Vibro: Video Browsing with Semantic and Visual Image Embeddings
29th International Conference on MultiMedia Modeling (MMM 2023)
TL;DR
Vibro represents a powerful tool for interactive video retrieval and browsing and is the winner of the Video Browser Showdown 2022. In this iteration the text-to-image & image-to-image search quality got improved by using general-purpose finetuning regime.
Abstract
Vibro represents a powerful tool for interactive video retrieval and browsing and is the winner of the Video Browser Showdown 2022. Following the saying of “never change a winning system” we did not change any of the underlying concepts nor added any new features. Instead, we focused on improving the three existing cornerstones of the software, which are text-to-image search, image-to-image search and browsing results with 2D sorted maps. The changes to these three parts are summarized in this paper, and in addition, an overview of the AVS-mode of vibro is given.
BibTeX
If you use our work in your research, please cite our publication:
@inproceedings{10.1007/978-3-031-27077-2_56,
author = {Schall, Konstantin and Hezel, Nico and Jung, Klaus and Barthel, Kai Uwe},
title = {Vibro: Video Browsing With Semantic And Visual Image Embeddings},
year = {2023},
isbn = {978-3-031-27076-5},
publisher = {Springer-Verlag},
address = {Berlin, Heidelberg},
url = {https://doi.org/10.1007/978-3-031-27077-2_56},
doi = {10.1007/978-3-031-27077-2_56},
booktitle = {MultiMedia Modeling: 29th International Conference, MMM 2023, Bergen, Norway, January 9–12, 2023, Proceedings, Part I},
pages = {665–670},
numpages = {6},
keywords = {Content-based video retrieval, Visual and textual co-embeddings, Visualization, Exploration, Image browsing},
location = {Bergen, Norway}
}