Scholarlink

Edge Intelligence deploys AI models directly on devices, reducing latency, bandwidth, and improving privacy. We implement lightweight image classification on an ESP32 for real-time pearl millet disease detection. Three compact CNNs—MobileNet, FOMO, and a custom CNN—are trained on a dataset collected from Coimbatore farms. The custom CNN achieves the highest accuracy (86%) and is deployed on an ESP32-CAM for offline inference, suitable for rural areas. This demonstrates a practical embedded-AI pipeline on low-cost hardware (

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Deploying Compact Image Classifier on Edge Device for Real-Time Pearl Millet Disease Detection

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • J. Chalmers,
  • J. Aravinth,
  • T Senthil Kumar,
  • I. Johnson

摘要

Edge Intelligence deploys AI models directly on devices, reducing latency, bandwidth, and improving privacy. We implement lightweight image classification on an ESP32 for real-time pearl millet disease detection. Three compact CNNs—MobileNet, FOMO, and a custom CNN—are trained on a dataset collected from Coimbatore farms. The custom CNN achieves the highest accuracy (86%) and is deployed on an ESP32-CAM for offline inference, suitable for rural areas. This demonstrates a practical embedded-AI pipeline on low-cost hardware (