1.There are two folders in Dataset-B and Dataset-C:
the ¡°Without data augmentation¡± folder contains the original, non-augmented datasets, while the ¡°Data augmentation¡± folder contains the augmented datasets.
In the augmented datasets, filenames that include ¡°aug¡± indicate augmented samples, whereas filenames without ¡°aug¡± correspond to the original samples.

2.For bulk dataset downloads, please contact:  liuyan926@mail.sdu.edu.cn

3.The file split.py is used for splitting the dataset into training, validation, and test sets.¡±



Python quick-start snippet£º

1.1 Recommended Environment

- Python ¡Ý 3.8 (recommended: Python 3.9)
- PyTorch ¡Ý 2.0
- CUDA ¡Ý 11.0 (optional, for GPU acceleration)
- :contentReference[oaicite:0]{index=0}
- numpy
- opencv-python
- matplotlib

1.2 Create Conda Environment (Recommended)
conda create -n yolo11 python=3.9 -y
conda activate yolo11


The following code snippet provides a minimal example for training the YOLO11 model using the Ultralytics framework:


import torch
from ultralytics import YOLO
import warning

warnings.filterwarnings('ignore')

def main():
    # Select device
    device = 0 if torch.cuda.is_available() else 'cpu'
    print(f"Using device: {device}")


    # Load model (custom YAML or pretrained weights)
    model = YOLO('ultralytics/cfg/models/11/yolo11.yaml')
    # model = YOLO('yolo11m.pt')  # alternative: load pretrained model

    # Dataset configuration file
    data_path = 'path/to/your/dataset.yaml'

    # Training parameters
    epochs = 500
    imgsz = 640
    batch = 8
    workers = 12

    # Start training
    model.train(
        data=data_path,
        epochs=epochs,
        imgsz=imgsz,
        batch=batch,
        workers=workers,
        device=device
    )

if __name__ == "__main__":
    main()

.