Pytomography

Latest version: v3.1.1

Safety actively analyzes 641082 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 4 of 6

1.0.0

* "image" has been renamed to "projection" everywhere. This is due to the fact that "image" is ambigious and can refer either to a 3D reconstruction (what clinicians call image) or a set of projections (what a mathematician would call image). As such, we now call the two spaces "object" space and "projection" space. The `projections` submodule has been renamed to `projectors` to remove any confusion with the projections as being a data type. Various arguments like `im2im_transforms` have been renamed to `proj2proj_transforms`. Please see the tutorials for how code has changed.
* readthedocs page has been updated with a link to the PyTomography paper on arxiv. Eventually this will be updated with the paper link on Medical Physics, once it's published there.

0.10.0

Implemented filtered back projection
Changes to opening SIMIND data
Callback now only takes in n_iter

0.9.3

Fix bug with `torch.set_default_device`

0.9.2

Fix bug where `torch.set_default_device` caused module import to fail, since this isnt in pytorch<2.0

0.9.1

**Patch**: Kornia now a required dependency
Fixed issues with 2DPET PSF Modeling

0.9.0

- update documentation
- `SystemMatrix` is now an abstract class which needs to be implemented for different modalities. SPECT uses `SPECTSystemMatrix` now. This is not back-compatible with old code.
- Back projection inside `SPECTSystemMatrix` cleaned up: no longer takes in prior

Page 4 of 6

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.