2023 Jan Update: Be sure to read any warnings in the sub-repo, especially the isles2017 folder.
2021 May Update: this repository has been restructured. Previous projects will be archived and no longer be updated.
Archived (no longer updated):
- isles2017. Deep learning for ischemic stroke lesion segmentation. The related paper is Enhancing the Extraction of Interpretable Information for Ischemic Stroke Imaging from Deep Neural Networks.
- Multiv. Image slice sampler.
- Attention-Gated-Networks_auxiliary. Auxiliary codes for Attention Gated Networks.
Figure 1. (left) An image input showing a castle. (right) The output of LRP to "explain" the decision of the algorithm to categorize the image into the correct class "castle". The images are obtained from the original website http://www.heatmapping.org/tutorial/.
Figure 2. (A) Raw LRP output. (B,C) LRP output with filters applied. (A2-C2) LRP output overlaid on a slice of RBF modality of an MRI scan. (D) Ground-truth lesion segmentation (E) Predicted lesion (F) A slice of RBF modality of an MRI scan.
Figure 3. Visualization of automated sampler.
For more details, visit each project folder.
Related links: