ITK 5.4.7: Expanded Image IO, Faster DTI, and Reliable Builds
Published on August 10, 2026 by Matthew McCormick
Weโre excited to announce ITK 5.4.7, the cross-platform, open-source toolkit for N-dimensional scientific image analysis!
๐ฅ Highlights
๐ ITK Python packages now use the Pool thread backend by default, better tuned for most workloads. TBB remains available.
๐งญ NRRD IO now supports 5D datasets, including time sequences of 3D vector, velocity, and color images.
๐ผ๏ธ TIFF IO adds signed and unsigned 32-bit integer support for 2D images and multi-page volumes.
โก Diffusion tensor reconstruction is substantially faster by computing its constant tensor-basis SVD once instead of once per voxel.
๐ ๏ธ Single-precision FFTW builds, GDCM/DCMTK integration, CI caching, CastXML wrapping, and C17/C23 support are more robust.
๐ค Community Thanks
Big cheers to Andras Lasso, Bradley Lowekamp, Csaba Pinter, Hans Johnson, James Butler, and the entire ITK community! ๐๐
๐ Whatโs Next?
The ITK 6 Release Candidate is coming, with modernized C++, Python wrapping, and toolkit architecture.
The ITK 6 Release Candidate is coming, with modernized C++, Python wrapping, and toolkit architecture.
Download ITK 5.4.7 and supercharge your image analysis! ๐
https://docs.itk.org/en/latest/releases/5.4.7.html
https://docs.itk.org/en/latest/releases/5.4.7.html
Enjoy ITK! ๐๐ฅณ