Strictly Causal Streaming Video Anomaly Detection with a Theoretically-Grounded State-Space Core

The paper presents a strictly causal streaming video anomaly detector that updates a fixed‑size state in constant time per frame, eliminating the need for clip buffering or lookahead. Its core is a diagonal linear state‑space recurrence with a decay gate, trained via self‑supervised next‑embedding prediction on a frozen visual backbone. The authors derive a closed‑form link between the recurrence’s decay spectrum and detection delay, validate on UCSD Ped2 and CUHK Avenue, and report real‑time latency on Apple M3 Pro hardware (≈0.75 ms per frame).

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

• rag
• efficiency
• benchmarks