- Rating
- 1
- Confidence
- 4
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
This work proposes NOVAD, a video‑free framework for video anomaly detection. It trains category‑specialized expert models solely on structured textual narratives generated by an LLM. A detector‑guided segmentation pipeline invokes video‑language models only on candidate abnormal segments to produce anomaly explanations. The strengths lie in the end‑to‑end video‑free training pipeline, structured event prompting and comprehensive empirical evaluations. Major weaknesses include persistent text‑video modality gap, uneven cross‑dataset generalization, limited architectural novelty and non‑negligible performance drops in certain settings.
Critical strengths
-
The anomaly detector does not require real target‑domain training videos, which tackles the practical difficulty of collecting and annotating rare real‑world anomaly footage.
-
Events are structured into four sequential phases: beginning, following, climax and resolution. Controlled sampling of locations, participants and objects diversifies synthetic textual training material.
Critical weaknesses
-
The novelty of this paper is limited. The mixture‑of‑experts backbone is directly adapted from GS‑MoE, and few original insights are contributed to core model design.
-
The proposed NOVAD underperforms TD‑VAD on several anomaly classes such as explosion and arson. Only shallow analysis is offered and no actionable fixes are proposed for these failure scenarios.
-
The proposed NOVAD outperforms TD‑VAD but still falls behind real‑video‑trained baselines, and no practical mitigation strategy is provided in the paper.
-
Qualitative outputs reveal obvious unsupported contextual hallucinations from the VLLM module, and the paper does not present any approach to ease this hallucination issue.
-
The reproduced baseline of TD-VAD's result deviates from the originally reported value, and this circumstance weakens the reliability of comparative experiments.
-
Cross‑dataset generalization delivers inconsistent performance, and the results degrade heavily under certain transfer setups, and the authors do not discuss any concrete improvement plans. Knowledge transfer from UCF‑Crime to XD‑Violence works well while the reverse direction yields poor results, so could you further explain the underlying root causes?
Questions for authors
see the above weakness