Hi,
I repeated your work with resnet18 and I was hoping to get better results but it did not happen.
I used a pretrained resnet18 that I attached for you. I extracted Caffe image features with res18 and then I decoded brain features with fastl2lir and finally, I did reconstruction and the reconstruction didn't do. Actually, reconstruction started with a lower error than vgg19 but after iteration 35, the error started to big. why???
I did reconstruction with vgg19 and my results were very close to your result. Now, I need your help.
Hi,
I repeated your work with resnet18 and I was hoping to get better results but it did not happen.
I used a pretrained resnet18 that I attached for you. I extracted Caffe image features with res18 and then I decoded brain features with fastl2lir and finally, I did reconstruction and the reconstruction didn't do. Actually, reconstruction started with a lower error than vgg19 but after iteration 35, the error started to big. why???
I did reconstruction with vgg19 and my results were very close to your result. Now, I need your help.