Skip to content

Latest commit

 

History

History
26 lines (18 loc) · 820 Bytes

README.md

File metadata and controls

26 lines (18 loc) · 820 Bytes

!!! WARNING This repo is WIP and the README is usually outdated.

This projects hosts scientific python code for detection, segmentation and tracking of cells and nuclei in 2-D and 3-D microscopy images, denoising using StructN2V method, and detection of mitotic nuclei.

Basic usage

cd devseg_2/src
module load gcc/6.2.0
module load cuda/9.2.148
source myenv/bin/activate

## generate training data, train and predict on test data.
python e23_mauricio2.py 

Data is usually stored in ../expr/{file name}/{function name}/*. E.g. e23_mauricio2.train() stores data in ../expr/e23_mauricio2/train/*.

Run via SLURM with

sbatch -J e23-mau -p gpu --gres gpu:1 -n 1 -t  6:00:00 -c 1 --mem 128000 -o slurm_out/e23-mau.out -e slurm_err/e23-mau.out --wrap '/bin/time -v python e23_mauricio2.py'