Hello everyone!

I'm preparing a paper on one of the image processing conferences about development of GrowCut method. If someone could help me and tell these things:
1. How long does it usually take to segment one image with it? Does it become much faster with experience?
2. Some examples of typical images 'fed' to growcut and cases when it is most helpful.

Thanks for help in advance!

Hi Moroz

I will try to help you all I can.

But to be honest, it depends on the picture. Some pictures work first time and others I have given up and used other methods.

As you have already mentioned the program is not very good with hair and fine detail but with areas of good contrast it works great. (Better than PS7)

So, if you would like to post some ‘Test Images’ I would be glad to help.
If you want data on bigger images then you could post at ‘You send it’ or email them.


Ken

Hi Moroz

I will try to help you all I can.

But to be honest, it depends on the picture. Some pictures work first time and others I have given up and used other methods.

As you have already mentioned the program is not very good with hair and fine detail but with areas of good contrast it works great. (Better than PS7)

So, if you would like to post some ‘Test Images’ I would be glad to help.
If you want data on bigger images then you could post at ‘You send it’ or email them.
Ken

Thanks for your intention! :)

About test images - usually every researcher selects the test images by himself when publishing his method. What I would like to do - is to learn where people really use this thing, I mean to get a couple of images from real users.

I don't mean that I will certainly publish them (I won't if you don't want to) - the main goal is to see what GrowCut is usally used for.

2-3 examples like this - "this picture took me about 1 minute to cut out <smth>, worked Ok", "this picture I've worked for 30 minutes and gave up". This would be very helpful.
Actually what I would like is to