Preclinical Research Disclaimer
Strictly for preclinical small-animal (murine micro-CT) research use only. Not FDA-cleared or approved for human clinical, diagnostic, therapeutic, or radiotherapy treatment-planning use. Software and segmentation models are provided “as is”, without warranty of any kind.
100% Local & Air-Gapped Offline Privacy
All NIfTI decoding, WebAssembly neural network inference (onnxruntime-web v1.31.1, MIT), 3D post-processing, and WebGL surface rendering run locally inside your browser. No scan data is ever uploaded, and the app makes zero external CDN requests.
Upstream Datasets, Teacher Models & Libraries
- AutoMICE (Jiang et al., Bioengineering 2024;
10.3390/bioengineering11121255): Code and Hugging Face weights licensed Apache-2.0; Zenodo weights (10.5281/zenodo.13841139) licensed CC BY 4.0.
- TumSeg (Jensen et al., Sci. Data 2024;
10.1038/s41597-024-03814-y): ERDA micro-CT dataset (10.17894/UCPH.F7BCF864-BE18-4A16-95AD-6F22DEDB4265) licensed CC BY 4.0; KjaerLab/TumSeg pretrained weights and code licensed CC BY-NC 4.0.
- Rosenhain Micro-CT Atlas (Rosenhain et al., Sci. Data 2018;
10.1038/sdata.2018.294): Figshare imaging datasets licensed CC0-1.0, including the bundled demo scan sample/sample_mouse_ct.nii.gz (M01_024h, CC0-1.0).
- Training & Methodology Frameworks:
MONAI & SwinUNETR (Apache-2.0), nnU-Net (Apache-2.0).
- Bundled Runtime: Microsoft
onnxruntime-web v1.31.1 in vendor/ licensed under the MIT License (Copyright © Microsoft Corporation). All viewer, NIfTI, and WebGL code is first-party with zero third-party CDNs.