How to do image segmentation of grayscale images using UNet or other deep learning networks in MATLAB?
25 views (last 30 days)
Show older comments
KOUSHIK DAS
on 12 Sep 2022
Commented: KOUSHIK DAS
on 14 Oct 2022
I have a dataset of 1112 grayscale images. This dataset also contains the annotated images (groudn truth) which are mask images of RGB format. I have to do the image segmentation and segment these image pixels into 5 categories. Is It possible to do image segmentation of this dataset in MATLAB, if yes, then how? Should I use the grayscale images for this or only the annotated images?
Thank you.
0 Comments
Accepted Answer
Vignesh Murugavel
on 4 Oct 2022
It is possible to do image segmentation of images in MATLAB.
Refer to this Link to see the various options available: Image Segmentation
As you want to segment these image pixels into 5 categories , you could use either of the two types of image segmentation techniques depending on your need.
1 ) Semantic segmentation: In this Objects shown in an image are grouped based on defined categories. It treats multiple objects within a single category as one entity. For Example: A street would be segmented by “pedestrians,” “bikes,” “vehicles,” “sidewalks,” and so on.
You can also refer to this example which shows how to create and train a simple semantic segmentation
2) Instance segmentation: This is a refined version of semantic segmentation. It detects the instances of each category. For Example : Categories like “vehicles” are split into “cars,” “motorcycles,” “buses,” and so on.
You can refer to this Link to know more about how to get started: Instance Segmentation
You should use the grayscale images for the image segmentation algorithm and Ground truths to access the performance of image segmentation where each pixel of the image is compared with corresponding ground truth data.The objective is to minimize the error between the segmented image and ground truth information.
More Answers (0)
See Also
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!