How to count each pixel with new semantic image

I have done the following tutorial: Semantic Segmentation Using Deep Learning » Deep Learning - MATLAB & Simulink (mathworks.com)

I now want to count each pixel for when an image goes through the network and produces semantic data. How do i do this? when the tutorial does count each pixel in the tutorial it does this for one image and uses PXDs.

To perform the semantic segmentation for a random image te following code is used:

I = read(imdsTest);
C = semanticseg(I, net);
B = labeloverlay(I,C,'Colormap',cmap,'Transparency',0.4);
imshow(B)
pixelLabelColorbar(cmap, classes);

How can i use this data to count the pixels in this outputted image for the labels?

The code used at the start to coun each pixel is:

labelDir = fullfile(outputFolder,'labels');
pxds = pixelLabelDatastore(labelDir,classes,labelIDs);

Answer 1

[Varun Sai Alaparthi]

Hello Benjamin,

I understand that you are looking for a way to get the pixel count with respect to labels for output image but the output from which function wasn’t clearly mentioned.

However, If the output is from the ‘semanticseg’ function following code can be used

c = semanticseg(img,net);
out= reshape (c,1,[]);
Summary(out)

This will provide you the pixel count

Sincerely

Varun