Finally, let's move on and collect data from the masks and perform some simple data analysis.

regionprops

The main function that we will use to collect data from masks is called regionprops. This function is very versatile and can collect a number of different properties, such as the area, length, and width of each object, or the mean intensity.

The function syntax is regionprops(mask). Note that mask must be a logical matrix, with each pixel containing true (or 1) if it is part of an object in the corresponding image, and false (or 0) for the background.

As an example, let's use the same image file we used in the last module (demoBPAE.tif)

>> I = imread('demoBPAE.tif');
>> mask = imbinarize(I);
>> mask = imfill(mask, 'holes');
>> data = regionprops(mask)
data = 
  struct with fields:

           Area: 11307
       Centroid: [133.5671 92.0717]
    BoundingBox: [28.5000 51.5000 222 79]

By default, regionprops calculates the following properties:

Area of the object (equal to the number of true pixels in the mask)
Centroid or center-of-mass coordinate of the object
BoundingBox the coordinates of the smallest rectangle containing the object

Other properties

regionprops can measure a range of other properties. For a full list, have a look at the documentation (doc regionprops). Some common properties include

Area
Centroid
Circularity
MajorAxisLength
MinorAxisLength
Orientation

To get regionprops to measure the other properties, you should specify the list of properties as comma-separated strings. For example,

>> data = regionprops(mask, 'Area', 'Centroid', 'MajorAxisLength')

In other words, the regionprops syntax is regionprops(mask, properties), where the list of properties properties is optional.

A quick note on units

regionprops will return all units of length and area as pixels.