A new Retrieval Engine for vitrivr

Authors
Ralph Gasser, Rahel Arnold, Fynn Faber, Heiko Schuldt, Raphael WaltenspĆ¼l, Luca Rossetto
Type
In Proceedings
Date
2024/1
Appears in
MMM 2024 - The 30th International Conference on Multimedia Modeling
Location
Amsterdam, The Netherlands
Abstract

While the vitrivr stack has seen many changes in components over the years, its feature extraction and query processing engine traces its history back almost a decade. Some aspects of its architecture and operation are no longer current, limiting the entire stack’s applicability in various use cases. In this paper, we present the first glimpse into vitrivr’s next-generation retrieval engine and our plan to overcome previously identified limitations.

Comments

VBS 2024

 

This preprint has not undergone peer review (when applicable) or any post-submission improvements or corrections. The Version of Record of this contribution is published in MMM 2024 - The 30th International Conference on Multimedia Modelingand is available online at https://doi.org/10.1007/978-3-031-53302-0_28.