Repository containing 2 core datasets of fixed and live plant organs imaged with confocal and light sheet microscopes together with ground truth cell instance segmentation.
For more information about the data acquisition, ground truth creation, please refer to https://www.biorxiv.org/content/10.1101/2020.01.17.910562v1
The files are located in their respective components.
1. LateralRootPrimordia - light-sheet volumetric data with voxel size: (0.25x0.1625x0.1625 µm^3) (ZYX). Courtesy of Alexis Maizel lab, COS, Heidelberg University, Germany
2. Ovules - confocal volumetric stacks with voxel size: (0.235x0.075x0.075 µm^3) (ZYX). Courtesy of Kay Schneitz lab, School of Life Sciences, Technical University of Munich, Germany
3. InflorescenceMeristem - confocal volumetrics stacks (raw data, no labels). Courtesy of Jan Lohmann lab, COS, Heidelberg University, Germany
4. CLSMLeaf - confocal volumetrics stacks (raw data + surface meshes). Courtesy of Miltos Tsiantis, MPIPZ, Cologne, Germany
Training and evaluation of the segmentation results was done using PlantSeg software: https://github.com/hci-unihd/plant-seg
Each dataset consist of HDF5 files organized into train/validate/test sub-folders. Each HDF5 file contains 2 internal datasets:
1. 'raw' - raw image of type uint8
2. 'label' - ground truth labels of type uint16
Additionally, the HDF5 files in the training subset of the Ovules dataset contain 'label_with_ignore' dataset, where all unlabeled cells were assigned an ignore label of -1.
If you use this data in your research please cite:
@article{wolny2020accurate,
title={Accurate and versatile 3D segmentation of plant tissues at cellular resolution},
author={Wolny, Adrian and Cerrone, Lorenzo and Vijayan, Athul and Tofanelli, Rachele and Barro, Amaya Vilches and Louveaux, Marion and Wenzl, Christian and Strauss, S{\"o}ren and Wilson-S{\'a}nchez, David and Lymbouridou, Rena and others},
journal={Elife},
volume={9},
pages={e57613},
year={2020},
publisher={eLife Sciences Publications Limited}
}
