轻量级联邦学习框架提升弱监督视频异常检测效果
Parser-Free VLM Verification for Federated Weakly Supervised Video Anomaly Detection
- 仅在客户端训练紧凑的MIL评分器，用冻结的VLM后验验证可疑片段
- 基于词元概率的连续异常得分使帧级AUC与AP均超越基线
- 无需解析文本或时序后处理，适合资源受限的分布式场景
当监控数据分布存储、标签弱且资源受限时，视觉语言模型如何助力视频异常检测？现有弱监督方法多依赖集中式训练，基于VLM的扩展又需密集推理、生成解释或额外适配。本文提出轻量级联邦MIL-VLM级联框架：仅在客户端训练紧凑的MIL评分器，而冻结的VLM在事后验证高分可疑片段。研究两种反馈接口：解析生成文本决策与基于词元概率的连续异常得分接口。在UCF-Crime数据集上，使用InternVL3.5-2B和Qwen3-VL-2B-Instruct测试表明，文本生成验证经时序后处理可提升帧级AUC，但对提示、解析器、模型选择及平滑敏感；而词元概率接口提供无需解析的稳定信号，在两个VLM上均显著提升帧级AUC与AP，且主配置无需时序后处理。因可疑片段可独立更新，该接口为文本生成验证提供简洁的局部替代方案。
原文摘要 · Abstract (English)
How can vision-language models help video anomaly detection (VAD) when surveillance data remain distributed, weakly labeled, and resource-constrained? Most weakly supervised VAD methods assume centralized training; recent VLM-based extensions further rely on dense inference, generated explanations, or additional adaptation. We introduce a lightweight federated MIL-VLM cascade in which only a compact MIL scorer is trained across clients, while a frozen VLM verifies high-scoring suspect segments post hoc. We study two VLM feedback interfaces: parsed text-generation decisions and a logit-based interface that extracts a continuous anomaly score from next-token Yes/No probabilities. Experiments on UCF-Crime with InternVL3.5-2B and Qwen3-VL-2B-Instruct show that text-generation verification can improve frame-level AUC after diagnostic temporal post-processing, but remains sensitive to prompts, parsers, model choice, and smoothing. In contrast, the logit interface provides a fixed parser-free signal that improves both frame-level AUC and frame-level AP over the MIL baseline across both VLMs, without temporal post-processing in its main configuration. Since suspect segments are updated independently once available, next-token logit feedback provides a simple segment-local alternative to text-generation verification.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。