Skip to content

AI & ML · Research experiment

Crop Disease Concept Net

An interpretable crop disease model that connects leaf images to six visible symptom concepts.

PythonResNet18Transfer learning
Explore repository
Crop Disease Concept Net project cover

The project

Transfer learning with a Concept Bottleneck Network: a pretrained ResNet18 feeds a six-symptom concept layer and a 14-class disease head. Training and prediction scripts support the crop disease dataset and synthetic demo data.

Repository activity ·

Inside the build

  • 01ResNet18 → symptom concepts → disease classification
  • 02Six concepts including rust patches, yellowing, and healthy tissue
  • 03Dataset split variants and a two-class comparison
  • 04Training, prediction, and synthetic demo-data scripts

Explore more

ReliefStream

Centralizes flood rescue requests with AI extraction, deduplication, a map, and volunteer coordination.

DocPilot RAG

Documentation Q&A combining keyword and dense retrieval, cross-encoder reranking, and source citations.