fix(core): return CSS pixels from the CPU worldToCanvas/canvasToWorld

[wayfarer3130]

The CPU transform built by calculateTransform is anchored on the canvas backing store, which getOrCreateCanvas sizes in device pixels, so pixelToCanvas and canvasToPixel speak device pixels. StackViewport passed their results straight through worldToCanvasCPU and canvasToWorldCPU, making the CPU path the only viewport path whose public coordinates are not CSS pixels: the GPU implementations divide by devicePixelRatio before returning, and the tools event layer derives its canvas points from getBoundingClientRect, which has no DPR term.

The two agree only at devicePixelRatio 1, which is why this went unnoticed. At any other ratio - HiDPI display, zoomed browser, mobile device - interaction is mis-scaled by exactly that factor, most visibly as panning that moves the image devicePixelRatio times further than the pointer. The offset is permanent rather than transient, zero at the canvas centre and growing with distance from it.

Convert at the public boundary only. The conversion deliberately does not live inside pixelToCanvas/canvasToPixel: internal rendering callers pass device pixels to those on purpose, and setToPixelCoordinateSystem applies the same transform to a 2D context where device pixels are the correct unit. The three internal round trips through canvasToPixel are bridged so their results are byte-for-byte unchanged.

resetCameraCPU already passed element.clientWidth/2 - CSS pixels - into canvasToWorldCPU, so it was silently wrong at DPR != 1 and is corrected by this change.

Reproduce before the fix: open a stack example with ?cpu=1, set browser zoom to 150%, and pan.

Fixes #2849

Note that testing this requires the touchAllTools PR to be merged first.

Context

Changes & Results

Testing

Checklist

PR

• [] My Pull Request title is descriptive, accurate and follows the
  semantic-release format and guidelines.

Code

• [] My code has been well-documented (function documentation, inline comments,
  etc.)

Public Documentation Updates

• [] The documentation page has been updated as necessary for any public API
  additions or removals.

Tested Environment

• [] "OS:
• [] "Node version:
• [] "Browser:

[coderabbitai]

📝 Walkthrough

Walkthrough

The CPU fallback now converts public CSS-pixel coordinates to device pixels for internal transforms and converts results back to CSS pixels. New helpers handle device-pixel ratios, and tests cover scaling, fallback behavior, and round trips.

Changes

CPU CSS Pixel Conversion

Layer / File(s)	Summary
Conversion helpers and validation packages/core/src/RenderingEngine/helpers/cpuFallback/rendering/cssPixelConversion.ts, packages/core/test/cssPixelConversion.jest.js	Added bidirectional CSS/device-pixel conversion using window.devicePixelRatio, with a fallback to 1. Tests cover integer and fractional ratios, missing ratios, identity conversions, and round trips.
StackViewport CPU coordinate boundaries packages/core/src/RenderingEngine/StackViewport.ts	Applied conversions at CPU world/index, focal-point, canvas/world, and world/canvas boundaries.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Possibly related PRs

• cornerstonejs/cornerstone3D#2769: Addresses DPR coordinate conversion in a different viewport implementation.
• cornerstonejs/cornerstone3D#2792: Adds related planar invariants and render-path parity coverage.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	The changes address issue #2849 by converting CPU viewport coordinates at the public boundary while preserving device-pixel internal rendering.
Out of Scope Changes check	✅ Passed	The conversion helpers, StackViewport changes, and tests are directly related to the linked CPU pixel-coordinate issue.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Title check	✅ Passed	The title clearly describes the CPU coordinate conversion fix and follows the required semantic-release format.
Description check	✅ Passed	The description explains the issue, implementation, impact, reproduction steps, and linked issue, but leaves the checklist and tested environment incomplete.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/browser-zoom-cpu

---

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.}

[wayfarer3130]

@jbocce - I've tested this on a tablet device (amazon fire) and it resolves hte scaling transform issue.

[jbocce]

Does the if(parallelScale) branch need to consider DPR as well?

[jbocce]

One small comment/question in the code changed, but I am also wondering why aren't other viewports affected by DPR or is it already handled?