Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/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.
https://github.com/KamitaniLab/DeepImageReconstruction
Last synced: about 2 months ago
JSON representation
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.
- Host: GitHub
- URL: https://github.com/KamitaniLab/DeepImageReconstruction
- Owner: KamitaniLab
- Created: 2017-12-26T02:37:06.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2023-07-03T09:57:13.000Z (about 1 year ago)
- Last Synced: 2024-07-09T04:38:21.905Z (3 months ago)
- Language: Python
- Homepage:
- Size: 286 KB
- Stars: 274
- Watchers: 27
- Forks: 60
- Open Issues: 6
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Deep Image Reconstruction
**Note: This demo code works with Python 2 and Caffe. Example code for the reconstruction with Python 3 + PyTorch is available at [brain-decoding-cookbook-public](https://github.com/KamitaniLab/brain-decoding-cookbook-public/tree/main/reconstruction).**
Data and demo code for [Shen, Horikawa, Majima, and Kamitani (2019) Deep image reconstruction from human brain activity. PLOS Computational Biology](https://doi.org/10.1371/journal.pcbi.1006633).
The preprint is availabe at bioRxiv ([Shen et al., 2017, Deep image reconstruction from human brain activity](https://www.biorxiv.org/content/early/2017/12/30/240317)).
## Dataset
- Raw fMRI data: [Deep Image Reconstruction@OpenNeuro](https://openneuro.org/datasets/ds001506)
- Preprocessed fMRI data, DNN features extracted from images, and decoded DNN features: [Deep Image Reconstruction@figshare](https://figshare.com/articles/Deep_Image_Reconstruction/7033577)
- Visual images: upon requeset via
## Code
### Requirements
- Python 2.7
- [icnn](https://github.com/KamitaniLab/icnn)
- Numpy
- Scipy
- Pillow (PIL)
- Caffe with up-convolutional layer
- https://github.com/dosovits/caffe-fr-chairs (Branch: deepsim)
- Both CPU and GPU installation are OK
### Usage
#### Preparation
1. Download data files from figshare (see [data/README.md](data/README.md)).
2. Download Caffe networks (see [net/README.md](net/README.md)).
#### DNN feature decoding from brain activity
You can skip the feature decoding from brain activity since we provide the decoded DNN features used in the original paper (see [data/README.md](data/README.md)).
We used the same methodology in our previous study for the DNN feature decoding ([Horikawa & Kamitani, 2017, Generic decoding of seen and imagined objects using hierarchical visual features, Nat Commun.](https://www.nature.com/articles/ncomms15037)).
The latest Python code for the DNN feature decoding is available at [GitHub:KamitaniLab/brain-decoding-cookbook-public](https://github.com/KamitaniLab/brain-decoding-cookbook-public).
#### Image reconstruction from decoded CNN features
We provide seven scripts that reproduce main figures in the original paper.
- 1_reconstruct_natural_image.py
- Reconstructing natural images from the CNN features decoded from the brain with deep generator network (DGN); reproducing results in Figure 2.
- 2_reconstruct_natural_image_without_DGN.py
- Reconstructing natural images from CNN features decoded from the brain without deep generator network (DGN); reproducing results in Figure 3A.
- 3_reconstruct_natural_image_different_combinations_of_CNN_layers.py
- Reconstructing natural images from CNN features decoded from the brain with different combinations of CNN layers; reproducing results in Figure 4.
- 4_reconstruct_shape_image.py
- Reconstructing colored artificial shapes from CNN features decoded from the brain; reproducing results in Figure 6A.
- 5_reconstruct_shape_image_different_ROI.py
- Reconstructing colored artificial shapes from CNN features decoded from multiple visual areas in the brain; reproducing results in Figure 7A.
- 6_reconstruct_alphabet_image.py
- Reconstructing alphabetical letters shapes from CNN features decoded from the brain; reproducing results in Figure 6B.
- 7_reconstruct_imagined_image.py
- Reconstructing imagined image from CNN features decoded from the brain; reproducing results in Figure 8.
PyTorch implementation of the reconstruction is available at [GitHub:KamitaniLab/brain-decoding-cookbook-public](https://github.com/KamitaniLab/brain-decoding-cookbook-public).
## Notes
### Enable back-propagation in the DNNs
In the demo code, we use pre-trained [VGG19](http://www.robots.ox.ac.uk/~vgg/software/very_deep/caffe/VGG_ILSVRC_19_layers.caffemodel) and pre-trained [deep generator network (DGN)](https://lmb.informatik.uni-freiburg.de/resources/binaries/arxiv2016_alexnet_inversion_with_gans/release_deepsim_v0.zip) ([Dosovitskiy & Brox, 2016, Generating Images with Perceptual Similarity Metrics based on Deep Networks. arXiv.](https://arxiv.org/abs/1602.02644)).
To enable make back-propagation, the following line should be added to the prototxt files (the file describes the configuration of the DNN):
```
force_backward: true
```
### Get DNN features before ReLU
In our study, we defined DNN features of conv layers or fc layers as the output immediately after the convolutional or fully-connected computation (i.e., before applying the Rectified-Linear-Unit (ReLU)).
However, as default setting of the pre-trained DNNs, ReLU operation is an in-place computation, which will override the DNN features we need.
To In order to use the DNN features before the ReLU operation, you need to modify the prototxt file as below (taking the VGG19 prototxt file as an example).
Original:
```
layers {
bottom: "conv1_1"
top: "conv1_1"
name: "relu1_1"
type: RELU
}
```
Modified:
```
layers {
bottom: "conv1_1"
top: "relu1_1"
name: "relu1_1"
type: RELU
}
```