ND Project is a research-oriented Python package for image processing, analysis, and segmentation.
It is designed to be modular, extendable, and compatible with PyTorch and NumPy, while also supporting custom operators and pipelines.
Requirements
- Minimal dependencies (`numpy`, `scipy`, `scikit-image`, etc.)
- PyTorch CPU version or GPU (Optional)
Features
- Modular operator system (differential operators, edge detectors, filters…)
- Compatible with both NumPy and PyTorch backends
- Advanced image analysis: texture descriptors (GLCM, entropy), denoising, segmentation
- Dataset generators for deep learning pipelines
- Logging and testing utilities