Open yxqlwl opened 7 years ago
I am an image processing student. I would like implement this for MRI images. i have these doubts after going through the work
It will be very useful for me if you can help me.
Thanks.
@zuie21 1.MatConvNet 2.Refer to DnCNN TrainingCodes 3.Using less layers is not related to overfitting I think. 4.Read the papper 5.feature descriptor? here?
are you a 医生吗
We are not getting much info from the read me note.
Regards, Suhail Aliyar
On 09-Feb-2018, at 2:29 PM, EmperorOfChina notifications@github.com wrote:
@zuie21 1.MatConvNet 2.Refer to DnCNN TrainingCodes 3.Using less layers is not related to overfitting I think. 4.Read the papper 5.feature descriptor? here?
are you a 医生吗
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.
Sorry, I'm confused. Is the training codes the same as the DnCNN? But DnCNN doesn't use HQS for training. The loss (Equation 10 in paper) is divided into 2 equations (Equation 7 and 8 in paper). Then we need to update z{k+1} by using matconvnet. I have no idea what loss should I choose when I wanna update the net. Besides, I don't know what Mathematical expression the Φ term is. Hope to get your help. Thanks a lot.
@XSLXANDY IRCNN is a model-based optimization method. It is not an end-to-end training method. IRCNN plugs the CNN denoisers into the HQS inference. So, you only need to train the denoisers.
Hi, sorry for interrupting. Could you please kindly provide the training code as well? Thank you~