Create README.md

README.md

@@ -0,0 +1,55 @@
+# AgriScan EfficientNetB0 — Crop and Disease Classification
+
+## Overview
+
+AgriScan EfficientNetB0 is an experimental agricultural image-classification system developed by CDSA (Cross Domain Solution Architect).
+
+The system uses two independently trained EfficientNetB0 classifiers to identify selected crops and predict their corresponding disease or healthy condition.
+
+## Model Architecture
+
+### 1. Crop Classifier
+
+Model file: `agriscan_efficientnet_crop.keras`
+
+Predicts one of three crop categories:
+- Maize
+- Potato
+- Tomato
+
+### 2. Disease Classifier
+
+Model file: `agriscan_efficientnet_disease.keras`
+
+Predicts one of 17 crop-disease or healthy classes across maize, potato and tomato.
+
+## Training
+
+- Architecture: EfficientNetB0
+- Framework: TensorFlow / Keras
+- Input dimensions: 224 × 224 pixels
+- Input format: RGB
+- Dataset: PlantVillage
+- Training method: Transfer learning followed by fine-tuning
+- Evaluation: Held-out test set from the source dataset
+
+## Evaluation
+
+Evaluation metrics and class-level reports are provided in the repository.
+
+Published performance figures should be taken from the completed evaluation artifacts, not assumed from the training process.
+
+## Intended Use
+
+This model is intended for research, prototyping and testing agricultural image-classification workflows.
+
+## Limitations
+
+- The classifiers cover only the crop and disease classes represented in training.
+- The model is not guaranteed to identify unfamiliar diseases or distinguish all non-plant images from crop leaves.
+- PlantVillage images may differ from photographs taken under real farming conditions.
+- Predictions should be validated using field data and, where necessary, agricultural expertise.
+
+## Development
+
+Developed as part of the AgriScan agricultural intelligence initiative by CDSA.