Segmentation training using original RGP image and binary mask

I have 2 datasets of just 1 object in several backgrounds taken by a smartphone camera.

Also, I have the masks for images in one of the datasets (created using Matlab image segmentrer).

Is there any suggestion of how to obtain mask for the other dataset ? based on the existing masks and rbg images ?

e.g., training a network in Matlab, built-in functions ?

Answer 1

[Image Analyst]

Try segnet: https://www.mathworks.com/help/vision/semantic-segmentation.html?s_tid=srchbrcm

It's possible you could use SURF: https://www.mathworks.com/help/vision/examples/object-detection-in-a-cluttered-scene-using-point-feature-matching.html

or even something simpler depending on how simple or complicated your object and backgrounds are.

[Ahmed Madhun]

Using segnet, i only see the possobility to manual label the images in the first database again rather that using the existing masks.