little-king/RiceBlast-Lesion-Segmentation

RiceBlast Lesion Segmentation Dataset(Preview)

This repository provides a preview subset of a rice leaf dataset, including instance-level leaf segmentation and semantic annotation of blast lesions. It is part of our ongoing research paper RBNet.

Status

• The associated manuscript is currently under peer review.
• The full dataset (including the complete image set, full annotations, and additional metadata) will be released upon paper acceptance.

File structure

.
├── data/coco/       # instance-level leaf segmentation datasets
│   ├── train2017/
│   │   └── xxx.jpg
│   ├── val2017/
│   │   └── xxx.jpg
│   ├── test2017/
│   │   └── xxx.jpg
│   └── annotations/
│       ├── instances_train2017.json
│       ├── instances_val2017.json
│       └── instances_test2017.json
│
└── datasets/        # semantic annotation of blast lesions
    ├── train/
    │   ├── images/
    │   │   └── xxx.png
    │   └── labels/
    │       └── xxx.png
    ├── val/
    │   ├── images/
    │   │   └── xxx.png
    │   └── labels/
    │       └── xxx.png
    └── test/
        ├── images/
        │   └── xxx.png
        └── labels/
            └── xxx.png

License

CC-BY-4.0

Citation

If you use this dataset in your research, please cite the corresponding paper:

X. Yang, C. Jin, et al., "An Instance–Semantic Cascade Framework with RBNet for In-Situ Rice Blast Severity Quantification in Multi-Leaf Field Scenes" (under review)

A BibTeX entry will be added after acceptance.

Contact

For questions, issues, or collaboration inquiries, please open an issue in this repository.