Handle constant B-spline mutual information inputs

[yinkev]

Closes #9018.

Description

B-spline global mutual information now remains finite for constant prediction or target tensors, rejects bin counts that cannot accommodate the required spline padding, and promotes half-precision spline distances before the cubic basis calculation.

Constant tensors have zero intensity range, so the previous bin width was zero and normalization produced NaNs. Separately, the cubic polynomial could overflow for float16 distances, while num_bins <= 4 made the padded interior denominator zero or negative.

Types of changes

• Non-breaking change (fix or new feature that would not break existing functionality).
• Breaking change (fix or new feature that would cause existing functionality to change).
• New tests added to cover the changes.
• Integration tests passed locally by running ./runtests.sh -f -u --net --coverage.
• Quick tests passed locally by running ./runtests.sh --quick --unittests --disttests.
• In-line docstrings updated.
• Documentation updated, tested make html command in the docs/ folder.

Validation

• float32 and float16 constant-input regressions fail on upstream/dev and pass with the fix;
• QUICKTEST=true python -m pytest tests/losses/image_dissimilarity/test_global_mutual_information_loss.py -q: 15 passed, 2 skipped;
• ruff check on both touched files;
• git diff --check.

[coderabbitai]

📝 Walkthrough

Walkthrough

GlobalMutualInformationLoss now requires more than four bins for the b-spline kernel. Its Parzen windowing path handles zero intensity ranges and casts float16 intermediate distances to float32. Tests cover invalid bin counts, constant full-precision inputs, constant half-precision inputs, finite losses, and finite gradients.

Estimated code review effort: 2 (Simple) | ~10 minutes

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Title check	✅ Passed	The title is concise and clearly reflects the main fix for constant B-spline mutual information inputs.
Description check	✅ Passed	The description matches the template well, with description, change types, and validation details included.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

🧹 Nitpick comments (1)

monai/losses/image_dissimilarity.py (1)

217-217: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Complete Google-style documentation for the added definitions.

• monai/losses/image_dissimilarity.py#L217-L217: add a Raises section for invalid bin counts.
• monai/losses/image_dissimilarity.py#L234-L235: document the b-spline-specific validation.
• tests/losses/image_dissimilarity/test_global_mutual_information_loss.py#L171-L196: add docstrings to the new test class and methods.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@monai/losses/image_dissimilarity.py` at line 217, The documentation for
GlobalMutualInformationLoss is incomplete: add a Google-style Raises section at
monai/losses/image_dissimilarity.py lines 217-217 covering invalid bin counts,
and document the b-spline-specific validation at lines 234-235. Add Google-style
docstrings to the new test class and each method in
tests/losses/image_dissimilarity/test_global_mutual_information_loss.py lines
171-196, describing the behavior each test verifies.

Source: Path instructions

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@monai/losses/image_dissimilarity.py`:
- Line 217: The documentation for GlobalMutualInformationLoss is incomplete: add
a Google-style Raises section at monai/losses/image_dissimilarity.py lines
217-217 covering invalid bin counts, and document the b-spline-specific
validation at lines 234-235. Add Google-style docstrings to the new test class
and each method in
tests/losses/image_dissimilarity/test_global_mutual_information_loss.py lines
171-196, describing the behavior each test verifies.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: ff583b21-a14a-43a8-85b2-78f57eb84094

📥 Commits

Reviewing files that changed from the base of the PR and between 3ee058b and dc3b7f0.

📒 Files selected for processing (2)

• monai/losses/image_dissimilarity.py
• tests/losses/image_dissimilarity/test_global_mutual_information_loss.py

[yinkev]

@coderabbitai review

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[yinkev]

Replaced by #9022 to conform to this repository's documented contributor workflow and branch-naming convention. The replacement uses the same signed commit and identical code diff: #9022