PlantdocAI ResNet50

PlantdocAI ResNet50 is an image classification model designed to identify plant diseases across 29 classes and 9 plant species. Developed using transfer learning, it employs a ResNet50 architecture with an ImageNet-pretrained, frozen backbone and a custom classification head. The model is positioned as the core component of PlantDoc AI, a mobile-first application that allows users, such as farmers, to photograph a leaf and receive a disease diagnosis along with offline voice advice in English, Urdu, or Sindhi.

The classifier itself is focused solely on predicting the disease class from an input image. Disease explanations, including information on causes, impacts, and precautions, are generated separately by a language model that only receives the predicted class name and does not process the image directly. This separation ensures that the image classification and explanation tasks are handled by distinct components within the broader PlantDoc AI system.

PlantdocAI ResNet50 is implemented using PyTorch and is tagged for use in agriculture, image classification, plant disease identification, and transfer learning applications. The model is available through Hugging Face and is distributed under the MIT license, allowing for open use and integration into other projects. Its intended use is for scenarios where plant disease classification is required, particularly in mobile contexts for agricultural support.