End-to-end deep image reconstruction from human brain activity

Kamitani Lab, Guohua Shen, Kshitij Dwivedi, Kei Majima, Tomoyasu Horikawa, Yukiyasu KamitaniFigshare 2019 · DOI · OpenAlex

Abstract

Shen, Dwivedi, Majima, Horikawa, and Kamitani (2019) End-to-end deep image reconstruction from human brain activity. Front. Comput. Neurosci. (bioRxiv preprint). <br><br>Here we provide pre-trained deep artificial neural networks (DNN) for reconstructing images from human brain activity. The shared pre-trained DNNs are the generator networks, which were trained to reconstruct the stimulus images from human brain fMRI activity. The pre-trained DNNs of 3 human subjects are shared here:<br><br>- trainedmodel_sub-01.zip<br>- trainedmodel_sub-02.zip<br>- trainedmodel_sub-03.zip<br><br>These pre-trained DNNs were trained for 500,000 mini-batch iterations, with batch size of 64. In each of the zip files, there are 2 files:<br><br>- generator.caffemodel: the Caffe caffemodel file for the trained generator<br>- generator_test.prototxt: the Caffe prototxt file for the generator<br><br>Demo code: GitHub
BibTeX
@inproceedings{lab2019end,
  author = {Kamitani Lab and Guohua Shen and Kshitij Dwivedi and Kei Majima and Tomoyasu Horikawa and Yukiyasu Kamitani},
  title = {{End-to-end deep image reconstruction from human brain activity}},
  booktitle = {Figshare},
  year = {2019},
  doi = {10.6084/m9.figshare.7916144.v1},
}

Loading related content...

Figure