This project contains several stardist models trained to segment DAPI-labeled nuclei in 2 and 3D, and barcodes in 3D.
These models are contained within the stardist_models.zip file and are described below.
### DAPI 2D
```sh
Model: stardist 2D
Training set: 3 ROIs, DAPI 2D embryos nc14, deconvolved, presegmented using image analysis.
Name: DAPI_2D_stardist_nc14_nrays:64_epochs:40_grid:2
Comment: third network trial. Excellent performance.
```
### DAPI 3D
```sh
Model: stardist 3D
Training set: pancreas, deconvolved images.
Name: DAPI_3D_stardist_17032021_deconvolved
Comment: First network for tissues.
```
### Barcodes 3D
`Folder: /mnt/grey/DATA/users/marcnol/pyHiM_AI_models/networks/`
```sh
Model: stardist 3D
Training set: 132 barcode images 256x256, simulations of PSF.
Name: PSF_3D_stardist_20210618_simu_deconvolved_thresh_0_01
Comment: Third network. Excellent performance in validation tests. Some bright spots are still not perfectly segmented (very low percentage).
```
