用视觉语言联合推理破解多种验证码，逼近人类水平。
VIPER Strike: Defeating Visual Reasoning CAPTCHAs via Structured Vision-Language Inference
- 分步解析图像布局与问题语义，动态匹配目标位置。
- 在六家平台测试中最高达93.2%成功率，优于现有方法。
- 适合作为安全研究者评估验证码鲁棒性的工具。
视觉推理验证码（VRC）结合视觉场景与自然语言查询，需对物体、属性和空间关系进行组合推理，正被广泛用于抵御自动化机器人。现有求解器分为两类：以视觉为中心的方法依赖模板特定检测器，在新布局下失效；以推理为中心的方法使用大语言模型（LLM），但难以处理细粒度视觉感知。两者均缺乏应对异构部署的通用性。我们提出ViPer，一个统一攻击框架，融合结构化多对象视觉感知与自适应的LLM推理。ViPer解析视觉布局，将属性关联至问题语义，并在模块化流程中推断目标坐标。在六个主流VRC提供商（VTT、Geetest、NetEase、Dingxiang、Shumei、Xiaodun）上评估，成功率达93.2%，接近人类水平。相比先前方法（GraphNet 83.2%、Oedipus 65.8%、Holistic 89.5%），ViPer持续领先。该框架在不同LLM后端（GPT、Grok、DeepSeek、Kimi）下保持90%以上准确率。为进一步预判防御，我们引入模板空间随机化（TSR），轻微扰动语言模板而不改变任务语义，显著降低求解器性能。该设计为构建人类可解但机器难破的验证码提供新方向。
原文摘要 · Abstract (English)
Visual Reasoning CAPTCHAs (VRCs) combine visual scenes with natural-language queries that demand compositional inference over objects, attributes, and spatial relations. They are increasingly deployed as a primary defense against automated bots. Existing solvers fall into two paradigms: vision-centric, which rely on template-specific detectors but fail on novel layouts, and reasoning-centric, which leverage LLMs but struggle with fine-grained visual perception. Both lack the generality needed to handle heterogeneous VRC deployments. We present ViPer, a unified attack framework that integrates structured multi-object visual perception with adaptive LLM-based reasoning. ViPer parses visual layouts, grounds attributes to question semantics, and infers target coordinates within a modular pipeline. Evaluated on six major VRC providers (VTT, Geetest, NetEase, Dingxiang, Shumei, Xiaodun), ViPer achieves up to 93.2% success, approaching human-level performance across multiple benchmarks. Compared to prior solvers, GraphNet (83.2%), Oedipus (65.8%), and the Holistic approach (89.5%), ViPer consistently outperforms all baselines. The framework further maintains robustness across alternative LLM backbones (GPT, Grok, DeepSeek, Kimi), sustaining accuracy above 90%. To anticipate defense, we further introduce Template-Space Randomization (TSR), a lightweight strategy that perturbs linguistic templates without altering task semantics. TSR measurably reduces solver (i.e., attacker) performance. Our proposed design suggests directions for human-solvable but machine-resistant CAPTCHAs.
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。