Laboratoire RDP » Biophysics Team » FLORIVAR

*Table of contents*

{{TOC}}

h1. image2geometry

Pipeline to detect, orient and measure sepals. The pipeline is based on different type of scripts that we developed and can be found in this repository.

- image type transformations are done with "ImageJ":https://imagej.net/Welcome macros

- python scripts making use of "MorphoGraphX":https://morphographx.org/

- python scripts that make use of "TimageTk":https://mosaic.gitlabpages.inria.fr/timagetk/

h2. 1. Preliminaries

You will need to *change different image file formats to .tif format*. To achieve this, you have suitable ImageJ macros in this repository's _bin/ImageJ_ folder : lif2tif, lsm2tif, inr2tif, tif2tif.

*Install the macro*: ImageJ → Plugins → Macros → Install → macroname.ijm

*Usage*: Their usage is the same. Lounch the macro and select the directory, where the files to transform are ; a new directory with the output tif files is created.

h2. 2. Normalise images

<pre> normalise.py imagefolder 30 </pre>

imagefolder=folder containing the images to normalise (.tif format)

30=common Otsu value

h2. 3. Detect the contour

The followings are different possibilities. Choose your favorite detection method and apply it on your normalised images.

h3. 3.1. By a threshold

h3. 3.2. Using the EdgeDetect of MorphoGraphX (mgx)

Use the MGX_EdgeDetect.py script on MorphoGraphX:

- open the script file with a texteditor;

- put the path to the directory of grey sepal images (line 13)

- save the script file

- drag and drop the script file on the MorphoGraphX black window

Remark: Before you lounch the scrcipt the first time in a session, you need to do a click-combinations on MorphographX (-> Process -> System -> set current stack -> Both + double click on Set current stack)

h3. 3.3. By the level set method (lsm)

cd in the directory with the smoothed tif files

<pre>

lsm-contour_mp.py imagefolder

</pre>

h3. 3.4. Lounch lsm on an already prepared initial contour (mgx for instance)

cd the directory, which contains a directory of (normalised and smoothed) images and another with the initial contours (same filename cooresponding to the same image)

<pre>

lsm-contour_mp-from-init.py imagefolder initcontourfolder

</pre>

detected contours are written in the folder « contours »

h2. 4. Orient sepals

cd in the folder containing the contours folder and type

<pre>orient_sepals.sh contours filetype</pre>

contours = folder name containing the contours

filetype = '.tif' or '.inr.gz'

the oriented sepals are written in the "contours_oriented" folder

h2. 5. Do measurements

h3. 5.1. All but area measurements

detected contours are in the folder « contours »

cd in the folder containing the oriented contours and type

<pre>measure_sepals.py foldername filetype</pre>

foldername = folder name containing the oriented contours

filetype = '.tif' or '.inr.gz'

the measurements are written in the "foldername_measures.csv" file

h3. 5.2. Area measurements

Use the MGX_CurvatureMapSegmentation.py script on MorphoGraphX:

- open the script file with a texteditor;

- put the path to the directory of sepal contours (line 13)

- save the script file

- drag and drop the script file on the MorphoGraphX black window

The segmented meshes (abaxial and adaxial surfaces are treated as cells) as well as one csv file per sepal containing the surface areas are saved.