digital confocal microscopy
├── LICENSE
├── README.md <- how to configure the repo
├── data
│ ├── external <- Data from third party sources.
│ ├── interim <- the data as the final input to models
│ ├── processed <- data after preprocessing
│ └── raw <- The original, immutable data dump.
│
├── docs <- A default Sphinx project; see sphinx-doc.org for details
│
├── models <- adopted models
├── models_weight <- trained weights
│
├── notebooks <- Jupyter notebooks for data inspection
│
├── references <- literature, papers and models
│
├── reports <- part of the results
│
├── requirements.txt <- The requirements file for reproducing the analysis environment, e.g.
│ generated with `pip freeze > requirements.txt`
│
├── setup.py <- makes project pip installable (pip install -e .) so src can be imported
│
└── tox.ini <- tox file with settings for running tox; see tox.readthedocs.io
the pipeline of the proejct works as below:
part of the results by segmentation are listed below:
when applied on the stack of the images, this model enables the virtual optical sectioning on widefiled microscopy. The segmented results reveal the 3D information of the targets.
the dataset could be found under this link, it contains two subsets: raw images from 15 stacks (each stack 20 slices) and its traditional segmented masks:
the original data form could also be found in /orignIMG: to give an example, this dataset only contains one stack of images. for more original stacks, please contact the author.
https://drive.google.com/drive/folders/1PUIGUsVCqTShgS81sLbgqc4u5NdTtZY4?usp=sharing
some of the training weights are also available in this folder, we offer two sets of the weights with depth 256 channels and 512 channels:
https://drive.google.com/drive/folders/1uA73P80NMypT3KOCk68oN3UIRpLaeL87?usp=sharing