A Benchmark Dataset for Paddy Bacterial, viral Disease and healthy Classification

Description

A Benchmark Dataset for Paddy Disease Classification is a labeled image dataset of paddy leaves, designed to support machine learning models in identifying and classifying crop diseases. It provides a standardized resource for benchmarking disease detection algorithms under diverse real-world conditions, encompassing Bacterial, viral infections, and healthy, along with healthy leaf samples

Steps to reproduce

Step 1: Collect Dataset Gather images of paddy leaves for all categories: Paddy Bacterial, viral diseases, and healthy Step 2: Label the Data Assign each image to its correct class and give labels: 0 (Healthy), 1 Paddy Bacterial, viral Disease (Bacterial Leaf Blight, Bacterial Leaf Streak, and Sheath Brown Rot (Bacterial); Rice Tungro Disease, Rice Yellow Mottle, Grassy Stunt, and Ragged Stunt Disease (Viral)) and healthy , Step 3: Organize the Dataset Create folders: train/ validation/ test/ Inside each, create subfolders for each class and place images accordingly. Step 4: Split the Dataset Divide the data into: 70% Training 15% Validation 15% Testing Step 5: Preprocess Images Resize images (e.g., 224×224) Normalize pixel values (0–1) Apply augmentation (flip, rotate, zoom) Step 6: Load the Dataset Use a framework like TensorFlow or PyTorch to load images from folders. Step 7: Choose a Model Select a model such as: CNN (custom) Pretrained models (ResNet, MobileNet, VGG) Step 8: Train the Model Train using the training set and validate using validation data. Step 9: Evaluate the Model Test the model using the test dataset and calculate: Accuracy Precision Recall F1-score Step 10: Reproduce Results Use the same: Dataset split Preprocessing steps Model parameters to achieve consistent results.

Institutions

• SR University

  Telangana, Warangal

Categories

Licence