KamitaniLab / brain-decoding-cookbook-public

MIT License
23 stars 5 forks source link

estimate cnn feat std #6

Closed fatemehkalantari1993 closed 1 year ago

fatemehkalantari1993 commented 1 year ago

Hi, How did you provide this file??

estimated_cnn_feat_std_VGG_ILSVRC_19_layers_ImgSize_224x224_chwise_dof1.mat

fatemehkalantari1993 commented 1 year ago

​I think I should use estimate_cnn_feat_std in utils.py for true images. in your code, you said that you used 10000 images for this task. what are these???

ShuntaroAoki commented 1 year ago

The file is included in VGG-19 model files. You can donwload them with the download script (https://github.com/KamitaniLab/brain-decoding-cookbook-public/blob/main/reconstruction/models/download.sh).

fatemehkalantari1993 commented 1 year ago

​I know, But I want to make this file for another network.

On Mon, 30 Mordad 1402 10:13 AM, Shuntaro Aoki @.***> wrote:

The file is included in VGG-19 model files. You can donwload them with the download script (https://github.com/KamitaniLab/brain-decoding-cookbook-public/blob/main/reconstruction/models/download.sh). — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

ShuntaroAoki commented 1 year ago

Ah, I see. I've uploaded the 10,000 images. Now you can find them at https://drive.google.com/file/d/1cJ-7zXs4rPS0ssJ3-F1X69REtSVN6Al5/view?usp=drive_link.

fatemehkalantari1993 commented 1 year ago

​Hi Thank you for your helpfirst, I extract features of 10000 images as 10000 .mat files in each layer and then I calculate the std of 10000 .mat files in each layer and finally, I calculate the mean of them in each layer. is this method true??? I just want to be sure this method is correct or not.

On Wed, 01 Shahrivar 1402 12:39 PM, Shuntaro Aoki @.***> wrote:

Closed #6 as completed. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>