Medical Image Analysis

The dataset contains the MR and model data for subjects S1, S2, and S3. For a detailed explanation of the files content, please refer to the included readme file.

3D meshes of the right nasal cavity extracted from publicly available head CTs obtained from the Cancer Imaging Archive (TCIA). The particular datasets are: 1. Data from head-neck cetuximab (http://doi.org/10.7937/K9/TCIA.2015.7AKGJUPZ) 2. Data from qin-headneck (http://doi.org/10.7937/K9/TCIA.2015.K0F5CGLI) The meshes are extracted using our segmentation methods.

Repository creation in progress. These data can be directly used to run the codes in https://github.com/juliacamps/Cardiac-Digital-Twin to generate and visualise digital twins and reproduce the results from "Harnessing 12-lead ECG and MRI data to personalise repolarisation profiles in cardiac digital twin models for enhanced virtual drug testing" (https://doi.org/10.1016/j.media.2024.103361). The supplement of the publication mentioned earlier contains additional information on the code and data structure.

# 👋 Faking_it team! BraTS submissions 🎬  ## 💡 Key Solutions (each subpage contains all the steps to reproduce the solutions): - **🥇 BraTS 2023 Task 1:** [Adult Glioma Segmentation](BraTS2023_Task1.md) - **🥇BraTS-ISBI 2024 GoAT:** [Generalizability Across Tumors Challenge](BraTS2024-ISBI_GoAT.md) - **🥇BraTS 2024 Task 1:** [Adult Glioma Post Treatment](BraTS2024_Task1.md) - **🥉BraTS 2024 Task 3:** [Meningioma Radiotherapy](BraTS2024_Task3.md) - **🏅BraTS 2024 Task 7:** [Synthesis (Global) - Missing MRI ](./BraTS2024_Task7.md)-> [Check out poster! ](./imgs/MICCAI2024-Poster-Task7_8.pdf) - **🥈BraTS 2024 Task 8:** [Synthesis (Local) - Inpainting](./BraTS2024_Task8.md) -> [Check out poster! ](./imgs/MICCAI2024-Poster-Task7_8.pdf) ✅ This repository contains the code and all the steps to reproduce the results of the submissions to BraTS 2023 Task 1, BraTS-ISBI 2024 GoAT, BraTS 2024 Tasks 1, 3, 7 and 8. ✅ Note that BraTS 2023 Task 1, BraTS-ISBI 2024 GoAT BraTS 2024 Tasks 1 and 3 are segmentation tasks and BraTS 2024 Tasks 7 and 8 are synthetic generation (using WDM 3D). ### :star_struck: We have released the trained weights! :partying_face: 💾 You can download them at ... You just need to place them in the correct place 🤓 ## Before running any experiments: 💻 For better experience, you should create a conda environment and have a machine with GPU. ### Segmentation tasks: ⚠️16GB of VRAM might be enough, however, we recomend using a GPU with 24GB. Be carefull with the amount of RAM you can use, as our code load the entire dataset to memory by default for faster training, but it might not be suitable for your machine. To reduce this, look into the data loaders. **💻 To create the conda environment:** 1. conda create -n BraTS_solutions python=3.11.9 2. pip install: 1. pip3 install torch torchvision torchaudio 2. pip install monai 3. pip install nilearn 4. pip install nibabel 5. pip install matplotlib 6. pip install pathlib 7. pip install einops 8. pip install tqdm 9. pip install SimpleITK 10. pip install nnunet 11. cd nnUNet_install 1. pip install -e . (nnunet v2) 12. cd mednext 1. pip install -e . (mednext) ### Synthetesis tasks: ⚠️ 40GB of VRAM is enough. We have set the `cache_rate=0` in `CacheDatase` in `c_bratsloader.py` file. For faster processing you can increase this number, up to 1. Be carefull with the amount of RAM you can use. 💻 To create the conda environment: 1. conda create --name wdm_submit python=3.10.1 2. pip install: 1. pip install nibabel 2. pip install monai 3. pip install blobfile 4. pip install PyWavelets 🤞 After running all commands, all dependencies should be installed. We performed our final tests on the 15 of October of 2024. If you find difficulties matching the versions, try to install the versions avaiable at that time. # If you find our work useful, please consider to ⭐️ **star this repository** and 📝 **cite our papers**: **BraTS 2023 Task 1:** [Adult Glioma Segmentation](BraTS2023_Task1.md) ``` @article{ferreira2024we, title={How we won BraTS 2023 Adult Glioma challenge? Just faking it! Enhanced Synthetic Data Augmentation and Model Ensemble for brain tumour segmentation}, author={Ferreira, Andr{\'e} and Solak, Naida and Li, Jianning and Dammann, Philipp and Kleesiek, Jens and Alves, Victor and Egger, Jan}, journal={arXiv preprint arXiv:2402.17317}, year={2024} } ``` **BraTS-ISBI 2024 GoAT:** [Generalizability Across Tumors Challenge](BraTS2024-ISBI_GoAT.md) ``` @inproceedings{ferreira2024generalisation, title={Generalisation of Segmentation Using Generative Adversarial Networks}, author={Ferreira, Andr{\'e} and Luijten, Gijs and Puladi, Behrus and Kleesiek, Jens and Alves, Victor and Egger, Jan}, booktitle={2024 IEEE International Symposium on Biomedical Imaging (ISBI)}, pages={1--4}, year={2024}, organization={IEEE} } ```  ---

FINTA/CINTA/GESTA/FIESTA datasets and neural network weights

Repository creation in progress. meta_data can be directly used to run the codes in https://github.com/juliacamps/Cardiac-Digital-Twin to generate and visualise digital twins and reproduce the results from "Harnessing 12-lead ECG and MRI data to personalise repolarisation profiles in cardiac digital twin models for enhanced virtual drug testing" (https://doi.org/10.1016/j.media.2024.103361). The supplement of the publication mentioned earlier contains additional information on the code and data structure. The monodomain simulations were performed using the configuration and mesh files in monodomain_monoalg3D_configuration_meshes.tar and using the version of monoAlg3D that can be found at https://github.com/bergolho/MonoAlg3D_C/tree/t-wave-personalisation-2024 The specific custom functions that were implemented in the t-wave-personalisation-2024 branch of the monoAlg3D code to enable the simulations can be found in monodomain_monoalg3D_custom_functions.tar.

This is the supplementary online material, including full data, evaluation, and executables, for the paper "Feature-based multi-resolution registration of immunostained serial sections" that appeared in Medical Image Analysis, Volume 35, January 2017, Pages 288–302. Same material was deposited online under https://gdv-server.inf.uni-bayreuth.de/gdvcloud/index.php/s/NnSov0O65n9Gp01 We also include here further supplementary files deposited at the journal page (http://www.sciencedirect.com/science/article/pii/S136184151630127X) under CC-BY licence. See there for the definite version of the paper or http://www.mathematik.uni-marburg.de/~lobachev/papers/lobachev-media16-registration-preprint.pdf for the preprint.

The MCO study whole slide image collection consists of 1500 digitised tissue slides of colorectal cancers. From 1994 to 2010, the Molecular and Cellular Oncology (MCO) Study group conducted a study of individuals undergoing treatment for colorectal cancer. For the study, they systematically collected tissue samples and clinical and pathological information from more than 1500 people who had tumours surgically removed from their large bowel. This collection represents one typical section from each tumour case, stained with Hematoxylin and eosin, and scanned using a x40 objective. The resolution of the digitised images approaches that visible under an optical microscope - more than 100,000 dpi. At this resolution, each image is around 2 Gigabytes, bringing the size of the 1500 images in the MCO Whole Slide Image Collection to 3 Terabytes. The MCO whole slide image collection is now available on the Intersect Australia Research Data Storage Infrastructure (RDSI) Node. Originating source(s): MCO research group, UNSW (1993-2011)