Datasets:
Tasks:
Image Classification
Modalities:
Image
Formats:
imagefolder
Languages:
English
Size:
1K - 10K
License:
| license: afl-3.0 | |
| task_categories: | |
| - image-classification | |
| language: | |
| - en | |
| tags: | |
| - disease | |
| - foliage | |
| - image | |
| - classification | |
| pretty_name: Foliagen | |
| size_categories: | |
| - 10K<n<100K | |
| datasets: | |
| - ASDID | |
| - PlantVillage | |
| - Soybean Leaf Disease Dataset | |
| # Foliagen Dataset | |
| The dataset is a foliage image dataset generated using a tool called [Foliagen](https://github.com/nabinpakka/FoliageGenerator). | |
| The generation process utilizes single leaf image datasets, namely, [ASDID](https://datadryad.org/dataset/doi:10.5061/dryad.41ns1rnj3), | |
| [PlantVillage](https://github.com/spMohanty/PlantVillage-Dataset), and [Soybean Leaf Disease Dataset](https://www.kaggle.com/datasets/sivm205/soybean-diseased-leaf-dataset). | |
| A generated foliage image dataset can be arbitrarily sized, with each of its image having a specified rate of diseased | |
| leaves (denoted by γ) and the rest being healthy ones. Being annotated by design, such generated datasets are valuable for (1) evaluating the SOTA classifiers when | |
| applied to practical use and (2) pre-training general SOTA classifiers, making it possible to fine-tune them using any real-world foliage image dataset for improved | |
| classification performance. | |
| ## License | |
| This dataset is released under the [Academic Free License v3.0 (AFL-3.0)](https://opensource.org/licenses/AFL-3.0) which is compatible with the license of the datasets it uses | |
| to generate foliage images. | |