Open-source codecs in Rust, with Apple Metal and NVIDIA CUDA acceleration.
Crates.io
01 / Applications
Use the codecs in your own application, or start with the Frames integrations for whole-slide imaging and machine learning.
Whole-slide imaging
Read scanner files and extract image regions with wsi-rs. It also reads supported Zeiss CZI brightfield scans compressed with JPEG XR. Export whole-slide images with wsi-dicom and convert QuPath annotations with wsi-dicom-annotations.
Image collections
Encode lossless JP2 or JPH images and decode smaller views for browsing. Transcode supported baseline JPEG inputs to HTJ2K from DCT coefficients without decoding to pixels.
DICOM workflows
Integrate JPEG, JPEG 2000 and HTJ2K pixel encoding and decoding into medical image applications. Choose lossless or lossy compression to match the image and transfer-syntax requirements.
Remote sensing
Read a tile, region or lower-resolution view of a large JPEG 2000 image. Use pixel access in mapping, earth observation and remote-sensing pipelines.
Training and inference
Decode JPEG or JPEG XR inputs, or feed JPEG 2000 and HTJ2K batches into Burn tensors. The MPSGraph integration keeps supported decoded batches on Apple GPUs.
Pixels for your application
Build thumbnails, tiled viewers and image-processing services with scaled and region decoding. Access native JPEG 2000 component planes for scientific image pipelines.
02 / Python
Use the Rust codecs from Python, with NumPy and PyTorch output. See the documentation for installation and supported inputs.
j2k
Decode supported JPEG 2000 and HTJ2K images into NumPy arrays or PyTorch tensors. GPU-enabled builds can decode supported inputs directly into PyTorch tensors on Apple GPUs or NVIDIA GPUs.
Installation
Build the j2k extension with Maturin in an existing Python environment. Enable Metal for Apple GPU output or CUDA for NVIDIA GPU output. The build instructions describe the platform requirements.
03 / Work with Frames
Get Frames working in your application. We offer integration help and custom development, including priority fixes for codec issues. If performance is the problem, we can profile your workload and work toward an agreed target.
Codec integration
Tell us what you’re building and where we can help. We can work with you on the integration and the codec changes your application needs.
Format support and performance
Bring us the file you’re having trouble with or the workload you want to improve. We can investigate it with you and develop a fix based on your images and software.
Send us a short description of your project. We’ll talk through what you need and work out how we can help.
04 / Public software
05 / Safety and GPU acceleration
The codec libraries expose safe public Rust APIs. Unsafe implementation code is confined to documented FFI, SIMD, allocation, and GPU boundaries.
Apple Metal and NVIDIA CUDA accelerate supported JPEG 2000 and HTJ2K operations. Supporting libraries also provide GPU paths for JPEG and JPEG XR. The J2K workspace’s CUDA kernels are written in Rust and compiled with cuda-oxide. CPU implementations are available on Linux, macOS and Windows. Coverage depends on the format, image and operation; see the J2K support matrix and JXR support guide.
Build image ingestion for AI, serve image regions, process geospatial imagery or integrate compression into your own application. Talk to Frames about integration and support for your application.
06 / Contact
Tell us what your software needs to do and where it is getting stuck.
Opens your email app with your message. You review and send it there.