Crop Disease Detection from Leaf Images with Lightweight CNNs
Project: KisanEye
Abstract
We build a smartphone-deployable convolutional network that detects 12 common crop diseases from leaf photographs. Using transfer learning on MobileNetV3 and a curated dataset of 24,000 field images, the model reaches 94.3% top-1 accuracy while running offline on entry-level Android phones.
No file attached to this sample paper.
Permalink: /paper/crop-disease-detection-from-leaf-images-with-lightweight-cnns-5dac5f