Synapse dataset

[janet-rgb]

I didn't get a competent train result and I have some questions regarding it.

I think it's because the dataset is not made of purely abdominal images, but instead include lower or higher parts of the body such as the heart or the testicle. The Groundtruth shows segmentation of 3 organs at most only when the given image is an image of the area consisting of organs such as the kidney, liver, etc. Otherwise, it's merely giving a black screen:

So I'm guessing it's because the data aren't purely abdominal images that it isn't doing so well ???

Or is it because I changed the max epoch to 10 and max iterations to 500?