[论文] VetClaw: An Edge-Cloud Multimodal Agentic System for Veterinary Diseas...

论文概要

研究领域: CV
作者: Syed Mhamudul Hasan, Anas AlSobeh, Hussein Zangoti, Abdur R. Shahid
发布时间: 2026-07-28
arXiv: 2607.26042

中文摘要

我们提出了VetClaw，一个用于早期兽医疾病筛查的边云多模态智能体系统。VetClaw使用摄像头模块作为边缘感知设备，将捕获的图像与可选的症状描述一起发送到服务器托管的视觉-语言模型进行零样本疾病分类。该系统将智能体交互与工作流编排分离：OpenClaw在边缘设备上提供调度、工具访问、用户交互和通知服务，而LangGraph管理有状态的筛查工作流，包括输入验证、图像传输、模型调用、安全检查、条件路由、故障处理和结构化日志记录。这种设计超越了静态图像分类，使系统能够收集视觉证据、调用外部模型、应用确定性安全规则并生成诊断支持警报。结果表明，仅图像的VLM预测仍然有限，而症状引导和多模态输入改善了零样本分类性能。因此，VetClaw将静态预测模型转变为可协调、安全感知、能使用工具、管理工作流、处理故障并升级不确定病例的系统。

原文摘要

We present VetClaw, an edge-cloud multimodal agentic system for early veterinary disease screening. VetClaw uses a camera module as an edge sensing device and sends captured images, together with optional symptom descriptions, to a server-hosted vision-language model for zero-shot disease classification. The system separates agent interaction from workflow orchestration: OpenClaw provides scheduling, tool access, user interaction, and notification services on the edge device, while LangGraph manages the stateful screening workflow, including input validation, image transmission, model invocation, safety checks, conditional routing, failure handling, and structured logging. This design moves beyond static image classification by enabling the system to collect visual evidence, invoke externa...

自动采集于 2026-07-30

#论文 #arXiv #CV #小凯