plant-pathology-2021

## Description
Dataset from the Plant Pathology 2021 (FGVC8) Challenge.

'
For Plant Pathology 2021-FGVC8, we have significantly increased the number of foliar disease images and added additional disease categories. This year’s dataset contains approximately 23,000 high-quality RGB images of apple foliar diseases, including a large expert-annotated disease dataset. This dataset reflects real field scenarios by representing non-homogeneous backgrounds of leaf images taken at different maturity stages and at different times of day under different focal camera settings.
'

The original dataset has one train split and a test split that was hidden for the challenge. I have taken 10% of train for a validation, using stratified sampling. I do not have access to the test samples.

- Website:
- https://www.kaggle.com/c/plant-pathology-2021-fgvc8
- https://sites.google.com/view/fgvc8/competitions/plantpathologychallenge2021

## Usage
This dataset is serving as a canonical example for multi-label image classificatino datasets with `timm`. The additions to train & val scripts for this are a WIP...

## Citation
```
Thapa, Ranjita, Zhang, Kai, Snavely, Noah, Belongie, Serge, and Khan, Awais. Plant Pathology 2021 - FGVC8.
https://kaggle.com/competitions/plant-pathology-2021-fgvc8, 2021. Kaggle.
```

## 数据集描述
本数据集源自植物病理学2021（FGVC8）挑战赛。

针对植物病理学2021-FGVC8赛事，我们大幅扩充了叶部病害图像的数量，并新增了病害类别。本年度数据集包含约23000张高质量苹果叶部病害RGB图像，其中包含一个规模可观的专家标注病害数据集。该数据集通过呈现不同成熟阶段、不同拍摄时段、不同相机对焦设置下的叶片图像非均匀背景，还原了真实的田间场景。

原始数据集包含一个训练划分集与一个赛事保密的测试划分集。本次处理中，我采用分层抽样方法从训练集中抽取10%作为验证集。我无法获取测试样本。

- 相关网址：
- Kaggle赛事页面：https://www.kaggle.com/c/plant-pathology-2021-fgvc8
- FGVC8官方赛事页面：https://sites.google.com/view/fgvc8/competitions/plantpathologychallenge2021

## 使用场景
本数据集作为`timm`框架下多标签图像分类数据集的标准范例。针对该数据集的训练与验证脚本扩充工作仍在进行中（WIP）。

## 引用格式

Thapa, Ranjita, Zhang, Kai, Snavely, Noah, Belongie, Serge, and Khan, Awais. 植物病理学2021 - FGVC8.
https://kaggle.com/competitions/plant-pathology-2021-fgvc8, 2021. Kaggle.

搜集汇总

背景与挑战

背景概述

该数据集来自Plant Pathology 2021 (FGVC8)挑战赛，包含约23,000张苹果叶部疾病的高质量RGB图像，涵盖多种疾病类别，图像背景非均质，拍摄于不同成熟阶段和不同时间，以反映真实田间场景，适用于多标签图像分类任务。数据集已划分为训练、验证和测试集（测试集隐藏），遵循Apache 2.0许可证，由timm维护，大小为16.09GB。