Apply semanticseg to multiple images

Hi, I am trying to perform semantic segmentation using Deep Learning, following the example in https://it.mathworks.com/help/vision/ug/semantic-segmentation-using-deep-learning.html.

I would like to use the network I trained to classify new images. However, I am able to perform semantic segmentation on a single images, but I want to process multiple images together and automatically label them with the network I trained.

C = semanticseg(I, net);

To this purpose, how could I apply the semanticseg function on multiple images of a folder?

Thank you to anyone who will reply to this question.

Answer 1

[Birju Patel]

To apply semanticseg to a images from a folder, you can pass in an imageDatastore to the semanticseg function:

imds = imageDatastore(folder);

pxds = semanticseg(imds, net);

See https://www.mathworks.com/help/vision/ref/semanticseg.html#d123e174246 for more information.

Answer 2

[yanqi liu]

yes，sir，may be make images to image dataset or just to 4-D format，such as

>> size(I3)

ans =

   720   960     3     2

>> C3 = semanticseg(I3, net);
>> size(C3)

ans =

   720   960     2

then we can get seg result，such as