cubic: CUDA-accelerated 3D Bioimage Computing

TL;DR

cubic addresses the growing scalability challenges in quantitative bioimage analysis for multidimensional microscopy by providing a device-agnostic Python library that transparently dispatches operations to GPU backends (CuPy and cuCIM) when available, while preserving CPU execution when not. It extends SciPy/scikit-image APIs with GPU-accelerated implementations, enabling end-to-end workflows from preprocessing to deconvolution and feature extraction in both 2D and 3D data. The authors benchmark cubic on 3D rescaling, 3D monolayer segmentation, and Richardson-Lucy deconvolution, reporting speedups from roughly 10× to more than 1000× with maintained fidelity and optional CPU fallback. The work emphasizes interoperability with the Python scientific stack, zero-copy data exchange with PyTorch, and open-source availability, positioning cubic as a scalable foundation for automated, high-throughput bioimage analysis.

Abstract

Quantitative analysis of multidimensional biological images is useful for understanding complex cellular phenotypes and accelerating advances in biomedical research. As modern microscopy generates ever-larger 2D and 3D datasets, existing computational approaches are increasingly limited by their scalability, efficiency, and integration with modern scientific computing workflows. Existing bioimage analysis tools often lack application programmable interfaces (APIs), do not support graphics processing unit (GPU) acceleration, lack broad 3D image processing capabilities, and/or have poor interoperability for compute-heavy workflows. Here, we introduce cubic, an open-source Python library that addresses these challenges by augmenting widely used SciPy and scikit-image APIs with GPU-accelerated alternatives from CuPy and RAPIDS cuCIM. cubic's API is device-agnostic and dispatches operations to GPU when data reside on the device and otherwise executes on CPU, seamlessly accelerating a broad range of image processing routines. This approach enables GPU acceleration of existing bioimage analysis workflows, from preprocessing to segmentation and feature extraction for 2D and 3D data. We evaluate cubic both by benchmarking individual operations and by reproducing existing deconvolution and segmentation pipelines, achieving substantial speedups while maintaining algorithmic fidelity. These advances establish a robust foundation for scalable, reproducible bioimage analysis that integrates with the broader Python scientific computing ecosystem, including other GPU-accelerated methods, enabling both interactive exploration and automated high-throughput analysis workflows. cubic is openly available at https://github$.$com/alxndrkalinin/cubic

Figures (3)

• Figure 1: GPU performance gains for 3D image rescaling operations. Results show speedup factors comparing GPU versus CPU execution across interpolation orders (0-5) for two sizes of 3D input images.
• Figure 2: CellProfiler 3D monolayer segmentation benchmark results. (a) Visual validation showing similar segmentation quality between CellProfiler and cubic implementations. (b) Performance comparison demonstrating significant computational speedup using GPU.
• Figure 3: Richardson-Lucy deconvolution of a large 3D image volume ($30\times2160\times2560$) with per-iteration image quality metrics tracking and CPU vs GPU performance comparison.