VADSK: Video Anomaly Detection with Structured Keywords
Thomas Foltz · 2025
Read VADSK: Video Anomaly Detection with Structured Keywords by Thomas Foltz (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-10653.txt.
Description
Thomas Foltz flags unusual events in surveillance video using weighted keywords in a 2025 study. Foundation models can describe any frame in words, and those descriptions generalize where pixel classifiers fail. The pipeline runs in two stages: first, descriptions of normal and anomalous frames are mined for the keywords that separate them, each given a weight. Then new frames are scored by their keyword profile through a small neural network.
Induction builds the vocabulary from examples, deduction applies it at speed. A keyword-based detector is light enough to run widely and easy to audit, since its reasons are words.
Catalog record
| CREATOR | Thomas Foltz |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-10653