PWD-Synthetic-Dataset

Dataset Overview

Dataset Name

• PWD‑Synthetic‑Dataset

Highlights

• Importance of Early Detection: Pine Wilt Disease is severe and lacks a cure, necessitating early detection.
• Dataset Advantage: Synthetic data creation surpasses traditional PWD data collection processes.
• Performance Boost: Merging real and synthetic data lifts the PWD F1 score to 92.88%.
• Broad Applicability: The synthetic‑data approach benefits forest protection and other agricultural fields.

Environment

• Python version: 3.11.4
• PyTorch version: 2.0.1
• GPU configuration: 8 × 2080Ti

Dataset and Pre‑trained Model Downloads

• Dataset Access: Obtain access by completing the form at this link.
• Pre‑trained Model: Download from this link.

Dataset Description

• Real Dataset
• Synthetic Dataset
• Synthetic Image Translation Dataset (I)
• Synthetic Image Translation Dataset (II)
• Synthetic Image Translation Dataset (III)

Training and Inference

• Training & Inference Code: Refer to and run this script.
• Test Set: Located in this folder.

Authors & Citation

• Authors: Yonghoon Jung, Sanghyun Byun, Bumsoo Kim, Sareer Ul Amin, Sanghyun Seo
• Citation:

@article{JUNG2024108690,
  title = {Harnessing synthetic data for enhanced detection of Pine Wilt Disease: An image classification approach},
  journal = {Computers and Electronics in Agriculture},
  volume = {218},
  pages = {108690},
  year = {2024},
  issn = {0168-1699},
  doi = {https://doi.org/10.1016/j.compag.2024.108690},
  url = {https://www.sciencedirect.com/science/article/pii/S0168169924000814},
  author = {Yonghoon Jung and Sanghyun Byun and Bumsoo Kim and Sareer {Ul Amin} and Sanghyun Seo}
}