rplatypus
Deep learning for 2D and 3D medical image segmentation and object detection in R.
Bridge R data pipelines and PyTorch architectures for DICOM and NIfTI files
rplatypus bridges the gap between R medical image handling packages and PyTorch deep learning models. It provides semantic and instance segmentation alongside object detection for formats like DICOM and NIfTI.
What it is
The package handles deep learning model specifications, fitting, and mask plotting directly from R using an isolated PyTorch engine. It features built-in support for architectures like U-Net and LinkNet with comparable evaluation metrics.
Who it's for
It suits R users working with medical imaging datasets who want to run deep learning without leaving their environment. It does not suit those who require pre-configured setups without downloading large Python dependencies.
Key features
Unified R Interface
Run deep learning pipelines entirely in R while computing via an isolated Python engine.
DICOM and NIfTI Support
Process standard medical imaging formats natively in 2D and 3D specifications.
Mask and Truth Plotting
Visualize predictions against ground truth masks with custom colormaps to inspect model errors.
Up and running in 3 steps
Our honest verdict
What's great
- Fills the gap in the CRAN Medical Image Analysis task view by bringing neural networks to R
- Automatically isolates and manages the underlying pyplatypus Python environment
- Provides visual diagnostics for predictions, truth, and missed regions
- Allows side-by-side comparison of different model specifications and loss functions
Where it falls short
- Requires downloading about 5 GB for the PyTorch wheel containing CUDA binaries
- Older GeForce GTX 10-series GPUs require a manual parameter call to avoid falling back to CPU
- Code written against the older 2020 release is incompatible with the current PyTorch rewrite
Pricing & license
The only cost is running it yourself, as the licence is not stated in the repository metadata.
Open remote AI-training roles
870 open remote AI-training roles, updated November 14, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Is Python required to be installed manually?
No, the package installs and isolates its pyplatypus engine automatically on the first call.
Will my old 2020 code work?
No, this is a complete rewrite running on PyTorch instead of TensorFlow.
