search

KamitaniLab / DeepImageReconstruction

Data and code for Shen, Horikawa, Majima, and Kamitani (2019) Deep image reconstruction from human brain activity. PLoS Comput. Biol. http://dx.doi.org/10.1371/journal.pcbi.1006633.
275 stars 61 forks source link

Another network #12

Open fatemehkalantari1993 opened 1 year ago

fatemehkalantari1993 commented 1 year ago

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.

ShuntaroAoki commented 1 year ago

Hi, I'm sorry for the delayed response. Could you share your ResNet decoded features, code (including the model definition and feature extraction code), as well as the pretrained model? This would allow us to check what is happening.

fatemehkalantari1993 commented 1 year ago

Hi, tanks a lot. I sent my information to "aoki.shuntaro.6n@kyoto-u.ac.jp". please check it.

fatemehkalantari1993 commented 1 year ago

Hi, Did you receive my data????

fatemehkalantari1993 commented 1 year ago

Hi, I am waiting for your reply. Thank you

fatemehkalantari1993 commented 1 year ago

Hi, This is my other email: nafisehproject@gmail.com n_kalantari@aut.ac.ir is my university email and It may not work well sometimes.