What do you mean by segmentation accuracy? Do you have the ground-truth? you shall do the difference or intersection of the segmentation and ground-truth; or something like that
If you have many images, then it is a lot of work, but it will be the more correct approach. You can do it in some specialised software, like photoshop, or other similar, or you can code an application for marking the object's contour points. What else you can do is to search for images databases that have segmentation groundtruth too
What do you mean by segmentation accuracy? Do you have the ground-truth? you shall do the difference or intersection of the segmentation and ground-truth; or something like that
How to create the ground-truth? I am planning to create manually. Is it a good decision?
If you have many images, then it is a lot of work, but it will be the more correct approach. You can do it in some specialised software, like photoshop, or other similar, or you can code an application for marking the object's contour points. What else you can do is to search for images databases that have segmentation groundtruth too
Thanks :) :)