Feature-driven Video Segmentation and Advanced Querying with vitrivr-engine

Authors
Luca Rossetto, Ralph Gasser
Type
Conference
Date
2025/1
Appears in
Proceedings of the 31st International Conference on Multimedia Modeling, MMM 2025
Location
Nara, Japan
Abstract

The vitrivr system has been a long-term contributor to the annual Video Browser Showdown. In our 2024 participation, we introduced a new core retrieval engine to the vitrivr stack, dubbed the vitrivr-engine. For the 2025 edition, we extend the engine with mechanisms for feature-driven video segmentation and additional querying means. We also extend the compatibility to additional storage backends, which enables vitrivr-engineto leverage advances made in the area of vector database systems.

Staff members
Research Projects