MONAI
MONAI copied to clipboard
Project-MONAI
AI Toolkit for Healthcare Imaging
Fixes #8656 Distance_transform_edt indices preallocation to use channel-first (C, spatial_dims, ...) layout for both torch/cuCIM and NumPy/SciPy paths, resolving “indices array has wrong shape” errors when return_indices=True. ### Description ```...
**Bug Description** Monai's distance_transform_edt allocates the feature-transform indices buffer with layout (ndim, C, …) instead of channel-first (C, spatial_dims, …), so SciPy/cuCIM reject the provided array and raise RuntimeError: indices...
Hi there! When using MaskedDice, my UNET doesn't train as well as when using the normal DICE. Upon closer inspection, I have noticed the following, which I believe to be...
Fixes #8627. Moves the perceptual loss code to MONAI repository https://github.com/Project-MONAI/perceptual-models and the checkpoints to Huggingface. ### Description This PR changes and simplifies the torch.hub loading process and gets the...
### Description This is an attempt to create a slim Docker image which is smaller than the current one to avoid running out of space during testing. Various fixes have...
Fixes #8619. ### Description As detailed in #8619, the `convert_to_channel_last` method of the `ImageWriter` class implicitly converts `data` from a `MetaTensor` to an `ndarray` [here](https://github.com/Project-MONAI/MONAI/blob/dev/monai/data/image_writer.py#L327). This PR addresses this issue...
### Description This PR addresses issue #8366 by implementing strict shape validation in `sliding_window_inference`. Per the feedback from maintainers (@ericspod), implicit guessing (heuristics) for channel-last data has been avoided. Instead,...
### Description This PR fixes the type hint for image reader to correctly handle Cupy arrays and removes the TODO. ### Types of changes - [x] Non-breaking change (fix or...
**Describe the bug** I have created some 4D NRRD files, in which the first dimension represents channels (different MRI scans). In my current use case, the 4D volumes are slices,...
Fixes #8396 . ### Description This PR fixes a bug where `Invertd` would fail with a RuntimeError when postprocessing pipelines contain invertible transforms (such as `Lambdad`) before `Invertd` is called....
