arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2512.04599cs.CV
收藏

通过视觉-语言分割融合进行恶意图像分析:一次检测、元素识别与定位

Malicious Image Analysis via Vision-Language Segmentation Fusion: Detection, Element, and Location in One-shot

Shanghai Jiao Tong University(上海交通大学);Microsoft Corporation(微软公司)

Sheng Hang, Chaoxiang He, Hongsheng Hu, Hanqing Hu, Bin Benjamin Zhu, Shi-Feng Sun, Dawu Gu, Shuo Wang

AI总结:

本文提出一种零样本方法,通过视觉-语言分割融合实现恶意图像的检测、元素识别与定位,提升细粒度审核的准确性和鲁棒性。

摘要:

检测非法视觉内容需要比图像级别NSFW标志更多;管理员必须了解哪些对象使图像非法以及这些对象的位置。我们介绍了一个零样本管道,能够同时(i)检测图像是否包含有害内容,(ii)识别每个关键元素并(iii)通过像素精确的掩码定位这些元素——所有这些都在一次通过中完成。系统首先应用基础分割模型(SAM)生成候选对象掩码,并将其细化为更大的独立区域。每个区域通过视觉-语言模型使用开放词汇提示进行评分,以评估恶意相关性;这些分数在融合步骤中产生一个汇总的恶意对象地图。多个分割器的集成增强了该管道对适应性攻击的抵抗力,这些攻击针对任何单一分割方法。在新注释的790张图像数据集上评估,该方法在元素级别召回率为85.8%,精度为78.1%,分割成功率为92.1%——在可比精度下,比直接零样本VLM定位方法的召回率高出27.4%。针对旨在破坏SAM和VLM的PGD对抗扰动,我们的方法的精度和召回率下降不超过10%,显示出对攻击的高鲁棒性。完整的管道可以在几秒钟内处理一张图像,无缝地融入现有的VLM工作流程,并构成了首个实用的细粒度、可解释的恶意图像审核工具。

英文摘要:

Detecting illicit visual content demands more than image-level NSFW flags; moderators must also know what objects make an image illegal and where those objects occur. We introduce a zero-shot pipeline that simultaneously (i) detects if an image contains harmful content, (ii) identifies each critical element involved, and (iii) localizes those elements with pixel-accurate masks - all in one pass. The system first applies foundation segmentation model (SAM) to generate candidate object masks and refines them into larger independent regions. Each region is scored for malicious relevance by a vision-language model using open-vocabulary prompts; these scores weight a fusion step that produces a consolidated malicious object map. An ensemble across multiple segmenters hardens the pipeline against adaptive attacks that target any single segmentation method. Evaluated on a newly-annotated 790-image dataset spanning drug, sexual, violent and extremist content, our method attains 85.8% element-level recall, 78.1% precision and a 92.1% segment-success rate - exceeding direct zero-shot VLM localization by 27.4% recall at comparable precision. Against PGD adversarial perturbations crafted to break SAM and VLM, our method's precision and recall decreased by no more than 10%, demonstrating high robustness against attacks. The full pipeline processes an image in seconds, plugs seamlessly into existing VLM workflows, and constitutes the first practical tool for fine-grained, explainable malicious-image moderation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑