Datasetpublic

marwankefah/nca-miccai2026-handson2

Neural Cellular Automata tutorial (MICCAI 2026), hands-on session 2 Code, preprocessed images and pre-trained weights for the notebook Robustness of Neural Cellular Automata under Distribution Shift. Path Content nca_lib.py, nca_widgets.py, nca_style.mplstyle, widgets/ session code and figure style bundle/bloodmnist_train64.npz BloodMNIST, 64 px, class-balanced training subset bundle/bloodmnist_test_multires.npz the same BloodMNIST test images at 64, 128 and 224… See the full description on the dataset page: https://huggingface.co/datasets/marwankefah/nca-miccai2026-handson2.

sourceHugging Facecc-by-sa-4.0updated 3d agoView on Hugging Face

0likes51downloads

Dataset Card

Neural Cellular Automata tutorial (MICCAI 2026), hands-on session 2

Code, preprocessed images and pre-trained weights for the notebook Robustness of Neural Cellular Automata under Distribution Shift.

Path	Content
`nca_lib.py`, `nca_widgets.py`, `nca_style.mplstyle`, `widgets/`	session code and figure style
`bundle/bloodmnist_train64.npz`	BloodMNIST, 64 px, class-balanced training subset
`bundle/bloodmnist_test_multires.npz`	the same BloodMNIST test images at 64, 128 and 224 px
`bundle/hippocampus_2d.npz`	MSD Task 04 Hippocampus, 2D slices at 64 px, split by patient
`checkpoints/*_seed0.pt`	pre-trained NCA, CNN and U-Net models

Sources and licences

—BloodMNIST, from MedMNIST v2 (Yang et al., Scientific Data 2023), CC BY 4.0. Original images: Acevedo et al., Data in Brief 2020.
—Hippocampus, from the Medical Segmentation Decathlon, Task 04 (Antonelli et al., Nature Communications 2022), CC BY-SA 4.0. The derived file inherits CC BY-SA 4.0, which is why this repository is released under that licence.

Methods

WBC-NCA (Deutges et al., MICCAI 2024); hNCA (Yang et al., IPMI 2025); Med-NCA (Kalkhof et al., IPMI 2023); NCA quality metric (Kalkhof and Mukhopadhyay, MICCAI 2023).