This video tutorial explains the process of fine tuning Detectron2 for instance segmentation using custom data. It walks you through the entire process, from annotating your data, to training a model, to segmenting images, to measuring object morphological parameters, to exporting individual masks (results) as images for further processing.

Code generated in the video can be downloaded from here: https://github.com/bnsreenu/python_for_microscopists/blob/master/330_Detectron2_Instance_3D_EM_Platelet.ipynb

All other code:
https://github.com/bnsreenu/python_for_microscopists

Detectron2 repo: https://github.com/facebookresearch/detectron2​

Annotations were done using Makesense: https://www.makesense.ai/

Dataset from: https://leapmanlab.github.io/dense-cell/
Direct link to the dataset: https://www.dropbox.com/s/68yclbraqq1diza/platelet_data_1219.zip

Data courtesy of:
Guay, M.D., Emam, Z.A.S., Anderson, A.B. et al. ​
Dense cellular segmentation for EM using 2D–3D neural network ensembles. Sci Rep 11, 2561 (2021). ​

Data annotated for 4 classes:
1: Cell
2: Mitochondria​
3: Alpha granule​
4: Canalicular vessel​

source