diff --git a/CHANGELOG.md b/CHANGELOG.md index d174df999f..36f423d6b1 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -5,6 +5,38 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/). ## [Unreleased] +## [1.6.2] - 2026-11-18 +### Addressed Security Advisories +* GHSA-f45g-w7qf-g6cv: Do not evaluate imports when reading an algorithm score (#9254) +* GHSA-987m-fcvp-pfxm: Treat BundleGen algos mapping values as literal data (#9253) +* GHSA-95v4-gm26-h6xf: Check nnUNet plans network names against allowed modules (#9237) +* GHSA-8wp8-g6wq-2m6r: Restore algo state values literally in algo_from_json (#9233) +* GHSA-jx6w-46x8-v3r5: Confine BundleAlgo.export_to_disk to the output directory (#9236) +* GHSA-9hf8-cr88-f7wg: Keep argument boundaries when building Auto3DSeg commands (#9234) +* GHSA-vq3w-pqfr-2w2r: Quote arguments in the NNI trial command (#9235) +* GHSA-4724-w94g-2rfp: Deprecate MMAR APIs and disable load_from_mmar (#9179) + +### Added +* Add ConvNeXt networks for 1D, 2D, and 3D inputs (#8995) +* Add MarchingCubes surface extraction and factor LPIPS into MONAI perceptual loss (#9107, #9287) +* Add average calibration losses, likelihood ratios, and diagnostic odds ratio metrics (#9092, #9105) +* Add per-component Surface Distance and optional native-resolution deep supervision for DynUNet (#8855, #9127) +* Add ROCm support for AMD Instinct GPUs and system resource logging with MLFlowHandler (#9153, #9052) + +### CI/Testing Fixes +* Expand ONNX export test coverage to 12 network architectures (#9073) +* Reduce test runtime by shrinking synthetic volumes, reusing test data, caching SwinUNETR SSL weights, and avoiding repeated mask allocation (#9034, #8985, #8987, #8993) +* Remove duplicate test cases (#9051) + +### Other Fixes +* Restrict `safe_eval` name resolution to caller-supplied variables (#9125) +* Fix diffusion inferers with diffusers-style schedulers and preserve device and dtype in `squarepulse` (#8870, #9005) +* Fix skipped `RandGridDistortiond` transforms, unspecified spatial sizes in resampling, compiled lazy resampling with `align_corners=True`, and `Spacingd` metadata updates (#8788, #9080, #9191, #9248, #9167) +* Fix Hausdorff percentile and ignored-voxel handling, per-component Dice, and EmbeddingCollapseMetric aggregation (#9096, #9033, #9195, #9171, #9134, #9117) +* Fix DatasetSummary empty-foreground handling, GridPatchDataset cache edge cases, CSVIterableDataset shuffling, and ZarrAvgMerger behavior with zarr 3 (#9043, #9101, #9249, #8816, #9077) +* Fix FID compatibility with scipy 1.18+ (#9076) +* Make `NormalizeIntensity` invertible and preserve non-contiguous inputs during inversion (#8905, #9183) +* Deprecate `hash_type` and MMAR APIs (#9130, #9179) ## [1.6.1] - 2026-09-25 ### Added diff --git a/CITATION.cff b/CITATION.cff index 6138d01bb4..e71f908ecf 100644 --- a/CITATION.cff +++ b/CITATION.cff @@ -6,8 +6,8 @@ title: "MONAI: Medical Open Network for AI" abstract: "AI Toolkit for Healthcare Imaging" authors: - name: "MONAI Consortium" -date-released: 2026-09-25 -version: "1.6.1" +date-released: 2026-11-18 +version: "1.6.2" identifiers: - description: "This DOI represents all versions of MONAI, and will always resolve to the latest one." type: doi diff --git a/docs/source/whatsnew.rst b/docs/source/whatsnew.rst index 8b8cec373e..9a95a52766 100644 --- a/docs/source/whatsnew.rst +++ b/docs/source/whatsnew.rst @@ -6,6 +6,7 @@ What's New .. toctree:: :maxdepth: 1 + whatsnew_1_6_2.md whatsnew_1_6_1.md whatsnew_1_6.md whatsnew_1_5_2.md diff --git a/docs/source/whatsnew_1_6_1.md b/docs/source/whatsnew_1_6_1.md index d9d47087c7..b81e4e05be 100644 --- a/docs/source/whatsnew_1_6_1.md +++ b/docs/source/whatsnew_1_6_1.md @@ -1,5 +1,5 @@ -# What's new in 1.6.1 🎉🎉 +# What's new in 1.6.1 This is primarily a security release with fixes for vulnerabilities identified through security alerts and tools. diff --git a/docs/source/whatsnew_1_6_2.md b/docs/source/whatsnew_1_6_2.md new file mode 100644 index 0000000000..baf9a2429a --- /dev/null +++ b/docs/source/whatsnew_1_6_2.md @@ -0,0 +1,25 @@ + +# What's new in 1.6.2 🎉🎉 + +This is primarily a security release with fixes for vulnerabilities identified through security alerts and tools. + +## Security Fixes + +A number of recent security alerts are addressed in this release: + +* GHSA-f45g-w7qf-g6cv: Do not evaluate imports when reading an algorithm score (#9254) +* GHSA-987m-fcvp-pfxm: Treat BundleGen algos mapping values as literal data (#9253) +* GHSA-95v4-gm26-h6xf: Check nnUNet plans network names against allowed modules (#9237) +* GHSA-8wp8-g6wq-2m6r: Restore algo state values literally in algo_from_json (#9233) +* GHSA-jx6w-46x8-v3r5: Confine BundleAlgo.export_to_disk to the output directory (#9236) +* GHSA-9hf8-cr88-f7wg: Keep argument boundaries when building Auto3DSeg commands (#9234) +* GHSA-vq3w-pqfr-2w2r: Quote arguments in the NNI trial command (#9235) +* GHSA-4724-w94g-2rfp: Deprecate MMAR APIs and disable load_from_mmar (#9179) + +Other additions include: + +* Add ConvNeXt networks for 1D, 2D, and 3D inputs (#8995) +* Add MarchingCubes surface extraction and factor LPIPS into MONAI perceptual loss (#9107, #9287) +* Add average calibration losses, likelihood ratios, and diagnostic odds ratio metrics (#9092, #9105) +* Add per-component Surface Distance and optional native-resolution deep supervision for DynUNet (#8855, #9127) +* Add ROCm support for AMD Instinct GPUs and system resource logging with MLFlowHandler (#9153, #9052)