Version 1.4.8

What's Changed

• Bugfix lesion_TPR_at_FPR by @joeranbosma in #20

Full Changelog: v1.4.7...v1.4.8

Version 1.4.7

What's Changed

• Update documentation permutation test (AI vs AI) by @joeranbosma in #11
• Evaluation progress bar by @joeranbosma in #13
• Implement lesion_TPR_at_FPR by @joeranbosma in #14
• Change image_utils.py to work with nnunetv2 predictions and ground truth labels with multi-class segmentations. by @NataliaAlves13 in #17
• Account for new keyword in sklearn v1.5+ by @joeranbosma in #19

New Contributors

• @NataliaAlves13 made their first contribution in #17

Full Changelog: v1.4.4...v1.4.7

Version 1.4.4

• Improve robustness of initialising Metrics (#7)
• Evaluate softmax predictions from command line (#9)
• Single threaded evaluation (#10)
• Documentation/cleanup

Version 1.4.1

Implemented linear sum assignment for lesion matching: Lesions candidates are matched to ground-truth lesions by optimising the number of matches first, and the total overlap second.

Version 1.3.2

Improved saving of metrics to JSON.

Version 1.3.1

Restructured lesion weights, for easy subset analysis of lesion-based metrics.

Version 1.2

Fixed calculation of average precision (AP) metric

Version 1.1.1

v1.1.1

Ensure submodule is initialised during setup

Version 1.0

Public release