Dataset for: Nutrispace: A Novel Color Space to Enhance Deep Learning Based Early Detection of Cucurbits Nutritional Deficiency

This dataset is an extension of our research titled "Nutrispace: A Novel Color Space to Enhance Deep Learning Based Early Detection of Cucurbits Nutritional Deficiency," which is currently under review. For utmost reproducibility, we've included a ZIP file and a Python file:

The ZIP file, "dataset_images," comprises 2700 segmented RGB images spanning 9 classes. The dataset is balanced and divided into train, validation, and test sets with a ratio of 0.70:0.15:0.15.

Class distribution of the dataset:
Crop | Class | Number of samples
---------------|------------------|-------------------
Ash gourd | Healthy | 300
Ash gourd | N deficiency | 300
Ash gourd | K deficiency | 300
---------------|------------------|-------------------
Bitter gourd | Healthy | 300
Bitter gourd | N deficiency | 300
Bitter gourd | K deficiency | 300
---------------|------------------|-------------------
Snake gourd | Healthy | 300
Snake gourd | N deficiency | 300
Snake gourd | K deficiency | 300
---------------|------------------|-------------------
| | Total = 2700

The converter.py file is a Python class file for transforming the RGB images to HSV, CIELAB, and our proposed color space - Nutrispace. The .py file contains a short documentation within it for the ease of use.

本数据集系我方研究项目“Nutrispace：一种旨在提升基于深度学习早期检测葫芦科植物营养缺乏的彩色空间”（目前正处于审稿阶段）的拓展。为确保极致的可复现性，我们特此附上了ZIP文件及Python脚本文件。

ZIP文件名为“dataset_images”，内含2700张分割后的RGB图像，共分为9个类别。数据集均衡划分，包括训练集、验证集和测试集，其比例为0.70:0.15:0.15。

数据集类别分布如下：

作物 | 类别 | 样本数量
---------------|------------------|-------------------
西葫芦 | 健康状态 | 300
西葫芦 | 氮缺乏 | 300
西葫芦 | 钾缺乏 | 300
---------------|------------------|-------------------
苦瓜 | 健康状态 | 300
苦瓜 | 氮缺乏 | 300
苦瓜 | 钾缺乏 | 300
---------------|------------------|-------------------
蛇瓜 | 健康状态 | 300
蛇瓜 | 氮缺乏 | 300
蛇瓜 | 钾缺乏 | 300
---------------|------------------|-------------------|
总计 | | 2700

“converter.py”文件为Python类文件，用于将RGB图像转换为HSV、CIELAB以及我们所提出的彩色空间——Nutrispace。该文件内含简要的文档说明，以便于用户使用。