Cells segmentation program made in MATLAB. This program is made for educational purpose. Analyse group of cells from pictures with image conversion, thresholding, filtering, regions separation and segmentation. The goal is to create video with the maximum of cells visible.

Original image	Image processed

Image regions	Image segmentation

A set of videos are available in the video folder. They show the different main step of the image processing.

First thing to do is to have MATLAB, you can find it below:
https://fr.mathworks.com/products/matlab.html

To run the project, you need to install the MATLAB packages:

• Image Processing Toolbox
• Signal Processing Toolbox
• Image Acquisition Toolbox

Once you have installed everything, you can start the project opening the root folder as a project and starting the main.m script in MATLAB.

The project is composed of one main script main.m that will call all the others script.

The other important scripts are :

• create_video_from_images.m to create video from a set of images
• image_processing.m to process images
• image_regions.m to display regions in a image
• image_segmentation.m to segment cells in a image

The three other following files are a set of functions for different way to convert an image to black and white (BW):

• im2bw_mean image to BW by using mean
• im2bw_median image to BW by using median
• im2bw_my_version image by manually changing threshold

Note

You can also use the Otsu's method to convert an image from grayscale to BW.

The pipe.m file is to use several functions on one variable For example, a conversion from grayscale to BW, processing and segmenting image. To use it there is four example in the code:

create_video_from_images(fps, inputFolder, imagesExtension, outputFolder, 'cells-4-bw-processed-segmented.avi', {im2bwSelectedFunc, @image_processing,@image_segmentation});

Usage:

variable = pipe(variable, {@func1,@func2,@func3});

MATLAB:
https://fr.mathworks.com/help/matlab/

Datacorner image processing course:
https://www.datacorner.fr/image-processing/

Quentin MOREL :

• @Im-Rises
• https://github.com/Im-Rises