This is an open source project dedicated to the segmentation and 3D reconstruction of plant organelles.
- Introduction
- Installation instructions
- Usage instructions
- Acknowledgements
- Pretrained checkpoint
- License
This a deep learning pipeline, called the organelle segmentation network (OrgSegNet), for pixel-wise segmentation to identify chloroplasts, mitochondria, nuclei, and vacuoles. With the help of OrgSegNet, you can complete a 3D reconstruction of a plant cell very quickly.
First, you need to complete the environment construction and dataset construction first. Please refer to Prepare Environment.md for installation and PrepareData.md for dataset preparation.
Great, I believe you have completed the environment setup and dataset build. Now let's train on the official dataset
We provide a training demo on jupyter notebook Train_OrgSegNet_demo, and a fine-tuning demo Fine-tune_OrgSegNet_demo, other demos can be found in file root "demo".
- System and Hardware requirements
- Test the dataset using the trained model(checkpoint)
- Instance generation based class activation map (CAM)
- Web implementation
- OrgSegNet code architecture and related guides
The pretrained checkpoint can be downloaded from the zonodo link.
As the official data set increases, the pre-trained model will be iterated gradually.
OrgSegNet is built on the MMCV and MMSegMentation open source frameworks. OrgSegNet was developed and maintained by Zeyu Yu, Institute of Agricultural Information Technology, School of Biosystems Engineering and Food Science, Zhejiang University.
Please cite the following paper when using OrgSegNet:
Feng, X., Yu, Z., Fang, H. et al. Plantorganelle Hunter is an effective deep-learning-based method for plant organelle phenotyping in electron microscopy. Nat. Plants (2023). https://doi.org/10.1038/s41477-023-01527-5