Abstract
Existing weakly supervised video violence detection (VVD) methods primarily rely on Euclidean representation learning, which often struggles to distinguish visually similar yet semantically distinct events due to limited hierarchical modeling and insufficient ambiguous training samples. To address this challenge, we propose PiercingEye, a novel dual-space learning framework that synergizes Euclidean and hyperbolic geometries to enhance discriminative feature representation. Specifically, PiercingEye introduces a layer-sensitive hyperbolic aggregation strategy with hyperbolic Dirichlet energy constraints to progressively model event hierarchies, and a cross-space attention mechanism to facilitate complementary feature interactions between Euclidean and hyperbolic spaces. Furthermore, to mitigate the scarcity of ambiguous samples, we leverage large language models to generate logic-guided ambiguous event descriptions, enabling explicit supervision through a hyperbolic vision-language contrastive loss that prioritizes high-confusion samples via dynamic similarity-aware weighting. Extensive experiments on XD-Violence and UCF-Crime benchmarks demonstrate that PiercingEye achieves state-of-the-art performance, with particularly strong results on a newly curated ambiguous event subset, validating its superior capability in fine-grained violence detection.
Abstract (translated)
现有的弱监督视频暴力检测（VVD）方法主要依赖于欧几里得表示学习，这在区分视觉上相似但语义不同的事件时往往存在困难，原因是层次化建模能力有限以及缺乏模糊训练样本。为了解决这一挑战，我们提出了PiercingEye，这是一种新颖的双空间学习框架，它结合了欧几里得几何和双曲几何以增强鉴别性特征表示。具体而言，PiercingEye引入了一种层敏感的双曲线聚合策略，并带有双曲线Dirichlet能量约束，用于逐步建模事件层次结构，以及一种跨空间注意力机制，有助于促进欧氏空间与双曲空间之间的互补特性交互。此外，为了缓解模糊样本稀缺的问题，我们利用大型语言模型生成基于逻辑指导的模糊事件描述，通过引入超球面视觉-语言对比损失进行明确监督，该损失优先处理高混淆样本，并采用动态相似度感知加权。在XD-Violence和UCF-Crime基准测试上的广泛实验表明，PiercingEye达到了最先进的性能水平，在新整理出来的模糊事件子集上尤其表现出色，验证了其在细粒度暴力检测方面的卓越能力。
URL
https://arxiv.org/abs/2504.18866
