Trouble labelling objects with semantic segmentation - pixelLabelColorbar

For my work, I am using a semantic segmentation network to look at photos and classify 3 objects: bubbles, water (the background), and a steel tube (where the bubbles come from). I have the following results:

I have been following this example on automated driving for the project:

https://www.mathworks.com/help/vision/examples/semantic-segmentation-using-deep-learning.html

My issue is that I cannot figure out how to implement the pixellabelcolorbar function or cmap variable in the example. I want to label the colors in my photos based on the classification like in the example. Can anyone please help with this?

The matlab live file is attached, relevant code is under "Running the Neural Network on Bubble Images" section. The test image is also attached.

The bubble classification network is here as a link (too large file size to attach): https://drive.google.com/open?id=1smBGJP3T8b47Cae5Wo6V3fvukuzo-76p

Any assistance is greatly appreciated

Answer 1

[Harsha Priya Daggubati]

Hi,

Considering the example you cited,i.e; https://www.mathworks.com/help/vision/examples/semantic-segmentation-using-deep-learning.html

pixelLabelColorbar adds the colorbar using the cmap(which is given by camvidColorMap function) and classnames while camvidColorMap function assigns a color to each class. Try replicating the same for your case.

Hope this clarifies your doubt!