middle-axis-estimations

This repository contains an algorithm for fitting a symmetric axis (w.r.t. the distance transform) between two points on an object in a 3D image volume.

Installation

• Clone this repo:

git clone git@github.com:uhlmanngroup/middle-axis-estimations.git

• Install dependencies via [conda]: conda env create -f environment.yml

Usage

Activate conda environment

conda activate ma_estimations

Processing

For processing the data make sure that all the files/image-volumes are in one folder

- data_directory
 - file1
 - file2
 - file3
 ...

Make sure that the endpoints-csv file contains the following fields and corresponding data entry for each file you want to process

File_name,Point,Position X,Position Y,Position Z,X_px_size_um,Y_px_size_um,Z_px_size_um,M,num_pole_pts

Example:

File_name,Point,Position X,Position Y,Position Z,X_px_size_um,Y_px_size_um,Z_px_size_um,M,num_pole_pts
C2-2020_02_07_E1_BM_thr_fill_iso,1,308.50,508.96,261.52,0.2306294,0.2306294,0.2306294,15,10
C2-2020_02_07_E1_BM_thr_fill_iso,2,757.60,594.40,293.50,0.2306294,0.2306294,0.2306294,15,10

The same applies for the midpoint-csv file which should have the following fields and data entries

File_name,Point,Position X,Position Y,Position Z,X_px_size_um,Y_px_size_um,Z_px_size_um,

Example:

File_name,Point,Position X,Position Y,Position Z,X_px_size_um,Y_px_size_um,Z_px_size_um,
C2-2020_02_07_E1_BM_thr_fill_iso,1,531,308,243,0.2306294,0.2306294,0.2306294,
C2-2020_02_07_E1_BM_thr_fill_iso,2,519,683,240,0.2306294,0.2306294,0.2306294,

For processing the data run:

python main.py --vol_dir path_to_directory_with_tif_files --endpoints_file path_to_csv_file_with_endpoints --midpoints_file path_to_csv_file_with_midpoints --zoom float --save_to path_to_save_file+name

Example:

python main.py --vol_dir 'example_data' --endpoints_file 'example_data/D3SPEG-RGD_endpoints.csv' --midpoints_file 'example_data/D3SPEG-RGD_midpoints.csv' --zoom 0.3 --save_to 'example_data/D3SPEG-RGD_descriptors.csv'