A large-scale optical microscopy image dataset of potato tuber for deep learning based plant cell assessment

This dataset includes 9,811 natural (unstained) and 6,127 stained brightfield microscopy images with three fold annotation such as physical, morphological and tissue grading based on weight, different section area and tissue zone respectively. The tuber weight ranging from 80-100 gm, 40-50 gm, 15-25 gm have been considered as large (L), medium (M), and small (S) respectively. Furthermore, the images have been labelled with L, M, and S to specify their respective weights. Further, the samples associated with each label have been numbered from 1 to 5, such as L1 refers to the first sample of a large tuber. For morphological grading, the tubers have been divided into three parts namely Z1, Z2, and Z3 which specify the bud, middle, and stem areas respectively. In tissue grading, the images have been captured from inner and outer core and labelled as IC and OC. We have used three types of staining agents such as Safranin O and Toluidine Blue O for visualizing cell walls and Lugols Iodine for distinguishing starch granules prominently. In addition, we prepared ground truth segmentation labels for three different potato tuber weights. The dataset contains three folders named as “stain”, “unstain” and “segmentation”. All the images are in JPG format. The raw microscopy images of potato tubers can be found in “stain” and “unstain” folder; whereas, the segmentation folder provides raw images with ground truth segmentation labels. The “stain” folder contains different stained images in the respective folders named as “safranin”, “toluidine” and “lugols”. The image labels information can be extracted from the image filenames itself. The file naming format for unstained image is as physical grading with sample no_ morphological grading_ tissue grading_ section no._ image no.; for example, M1_Z1_IC_Sec1_02.jpg refers to an image (first section out of five) taken from the inner core of Z1 of medium weight potato tuber (sample no 1). Similarly, the stained image file naming format is asphysical grading with sample no_ morphological grading_ tissue grading_ section no._ stain type_ image no.; for example, S1_Z2_OC_Sec3_lugol_02.jpg refers to an lugol’s iodine stained image (third section out of five) taken from the outer core of Z2 of small weight potato tuber (sample no 1). The segmentation folder contains two subfolders, named as “images” and “groundtruth”. The image files naming format is as physical grading_ image no.; for example, “L_2.jpg” represents the second image of a large potato tuber sample.

本数据集包含9811张未染色自然明场显微镜（brightfield microscopy）图像与6127张染色明场显微镜图像，采用三类标注方案：基于重量的物理分级、基于切片面积的形态学分级，以及基于组织区域的组织分级。其中，块茎重量区间80-100 gm、40-50 gm、15-25 gm分别划分为大（L）、中（M）、小（S）三个等级，并以L、M、S作为对应重量等级的标签。各等级标签对应的样本依次编号为1至5，例如L1代表大号块茎的第1份样本。

针对形态学分级，本数据集将块茎划分为Z1、Z2、Z3三个区域，分别对应芽部、中部与茎部区域。组织分级方面，图像采集自块茎的内芯与外芯区域，标注为IC（inner core）与OC（outer core）。本数据集使用三类染色剂：番红O（Safranin O）与甲苯胺蓝O（Toluidine Blue O）用于可视化细胞壁，卢戈氏碘液（Lugols Iodine）用于清晰区分淀粉颗粒。

此外，本数据集针对三种不同重量的马铃薯块茎制备了真值（ground truth）分割标注。数据集包含三个文件夹，分别命名为stain、unstain与segmentation。所有图像均采用JPG格式。马铃薯块茎的原始显微镜图像存放于stain与unstain文件夹中；而segmentation文件夹则提供带有真值分割标注的原始图像。stain文件夹下设safranin、toluidine与lugols三个子文件夹，分别存放对应类型的染色图像。

图像的标签信息可直接从图像文件名中提取。未染色图像的文件名格式为：物理分级_样本编号_形态学分级_组织分级_切片编号_图像编号，例如M1_Z1_IC_Sec1_02.jpg，代表取自中号块茎（样本1）Z1区域内芯的第1份切片的第2张图像（共5份切片）。染色图像的文件名格式为：物理分级_样本编号_形态学分级_组织分级_切片编号_染色类型_图像编号，例如S1_Z2_OC_Sec3_lugol_02.jpg，代表取自小号块茎（样本1）Z2区域外芯的第3份切片的卢戈氏碘染色图像（共5份切片）。

segmentation文件夹下设images与groundtruth两个子文件夹。其图像文件的命名格式为：物理分级_图像编号，例如"L_2.jpg"代表大号块茎样本的第2张图像。