Update SkiaSharp from 2.88.8 to 3.119.2

[Copilot]

• Update SkiaSharp version from 2.88.8 to 3.119.2 in eng/Versions.props
• Update MLImage.cs to fix breaking API changes:
  • Replace _image.GetPixelSpan() with GetPixels() pointer-based span with IntPtr.Zero validation
  • Replace SKFilterQuality.None with new SKSamplingOptions(SKFilterMode.Nearest) in ResizeFull()
  • Replace SKPaint.FilterQuality = SKFilterQuality.High with DrawImage + SKSamplingOptions(SKCubicResampler.Mitchell) in ResizeWithPadding() and ResizeWithCrop()
  • Remove obsolete SKPaint.FilterQuality from CloneWithGrayscale()
• Add explicit SkiaSharp package reference in csproj (no longer transitively pulled by native assets in 3.x)
• Fix csproj comment to accurately describe the SkiaSharp package references
• Build the ImageAnalytics project to verify compilation ✅

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Update to latest version of SkiaSharp</issue_title>
<issue_description>System Information (please complete the following information):

• OS & Version: Windows 10
• ML.NET Version: ML.NET v4.0.2
• .NET Version: .NET 8.0

Describe the bug

In https://github.com/dotnet/machinelearning-samples, open ObjectDetection project, explicitly depend on SkiaSharp package 3.116.1, run, and observe failure:

=====Identify the objects in the images=====

System.InvalidOperationException: Splitter/consolidator worker encountered exception while consuming source data
 ---> System.MissingMethodException: Method not found: 'System.ReadOnlySpan`1<Byte> SkiaSharp.SKBitmap.GetPixelSpan()'.
   at Microsoft.ML.Data.MLImage.get_Pixels()
   at Microsoft.ML.Transforms.Image.ImagePixelExtractingTransformer.Mapper.<>c__DisplayClass5_0`1.<GetGetterCore>b__1(VBuffer`1& dst)
   at Microsoft.ML.Transforms.Onnx.OnnxTransformer.Mapper.NamedOnnxValueGetterVec`1.GetNamedOnnxValueCore()
   at Microsoft.ML.Transforms.Onnx.OnnxTransformer.Mapper.NamedOnnxValueGetterVec`1.GetNamedOnnxValueKnownSize()
   at Microsoft.ML.Transforms.Onnx.OnnxTransformer.Mapper.NamedOnnxValueGetterVec`1.GetNamedOnnxValue()
   at Microsoft.ML.Transforms.Onnx.OnnxTransformer.Mapper.UpdateCacheIfNeeded(Int64 position, INamedOnnxValueGetter[] srcNamedOnnxValueGetters, List`1 activeOutputColNames, OnnxRuntimeOutputCacher outputCache)
   at Microsoft.ML.Transforms.Onnx.OnnxTransformer.Mapper.<>c__DisplayClass14_0`1.<MakeTensorGetter>b__0(VBuffer`1& dst)
   at Microsoft.ML.Data.DataViewUtils.Splitter.InPipe.Impl`1.Fill()
   at Microsoft.ML.Data.DataViewUtils.Splitter.<>c__DisplayClass7_1.<ConsolidateCore>b__2()
   --- End of inner exception stack trace ---
   at Microsoft.ML.Data.DataViewUtils.Splitter.Batch.SetAll(OutPipe[] pipes)
   at Microsoft.ML.Data.DataViewUtils.Splitter.Cursor.MoveNextCore()
   at Microsoft.ML.Data.RootCursorBase.MoveNext()
   at Microsoft.ML.Data.ColumnCursorExtensions.GetColumnArrayDirect[T](IDataView data, Int32 col)+MoveNext()
   at System.Linq.Enumerable.SelectEnumerableIterator`2.MoveNext()
   at System.Linq.Enumerable.TryGetElement[TSource](IEnumerable`1 source, Int32 index, TSource& element)
   at System.Linq.Enumerable.ElementAt[TSource](IEnumerable`1 source, Int32 index)
   at ObjectDection.Program.Main(String[] args) in C:\Projects\Intermediate\machinelearning-samples\samples\csharp\getting-started\DeepLearning_ObjectDetection_Onnx\ObjectDetectionConsoleApp\Program.cs:line 43
========= End of Process..Hit any Key ========

Root cause is ImageAnalytics.

To Reproduce
Steps to reproduce the behavior:

1. Download and use https://github.com/dotnet/machinelearning-samples
2. Open samples\csharp\getting-started\DeepLearning_ObjectDetection_Onnx\ObjectDetectionConsoleApp.ObjectDetection.csproj
3. Use NuGet to explicitly install SkiaSharp version 3.116.1
4. Run
5. Fail

Expected behavior

Shouldn't fail or are there alternatively/ways to configure?
Also the package dependency of ML.Net.ImageAnalytics claims SkiaSharp >2.88

Screenshots, Code, Sample Projects
See log error above.

Additional context
I am using ML.Net in a project that already depends on SkiaSharp explicitly.
</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes Update to latest version of SkiaSharp #7413

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[Copilot]

Pull request overview

Updates Microsoft.ML.ImageAnalytics to be compatible with SkiaSharp 3.x to prevent runtime MissingMethodException when users bring newer SkiaSharp versions into their apps.

Changes:

• Bumps SkiaSharp to 3.119.2 via eng/Versions.props.
• Adds an explicit SkiaSharp package reference to Microsoft.ML.ImageAnalytics.csproj.
• Migrates removed/obsolete APIs in MLImage (replaces GetPixelSpan() with GetPixels() pointer span; replaces SKFilterQuality with SKSamplingOptions / DrawImage).

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
src/Microsoft.ML.ImageAnalytics/Microsoft.ML.ImageAnalytics.csproj	Adds explicit managed SkiaSharp reference (3.x no longer comes transitively via native assets).
src/Microsoft.ML.ImageAnalytics/MLImage.cs	Updates pixel access and resize/grayscale code paths to SkiaSharp 3.x APIs.
eng/Versions.props	Bumps SkiaSharpVersion to 3.119.2.

Comments suppressed due to low confidence (1)

src/Microsoft.ML.ImageAnalytics/MLImage.cs:397

• Same as above: SKImage.FromBitmap(_image) in the crop path adds an extra allocation/native resource per resize. Prefer drawing from SKBitmap directly with SKSamplingOptions if the SkiaSharp 3.x API supports it, to reduce per-image overhead in pipelines.

            using SKCanvas canvas = new SKCanvas(dst);
            using SKImage image = SKImage.FromBitmap(_image);

            canvas.DrawImage(image, srcRect, destRect, new SKSamplingOptions(SKCubicResampler.Mitchell));

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[stephentoub]

@copilot address the feedback as is necessary / valuable

[Copilot]

@copilot address the feedback as is necessary / valuable

Addressed in 00d95ab:

• Fixed the inaccurate csproj comment about SkiaSharp native dependencies
• Added IntPtr.Zero / ByteCount validation in Pixels getter

The SKImage.FromBitmap suggestion isn't actionable — in SkiaSharp 3.x, DrawBitmap doesn't support SKSamplingOptions, so DrawImage (which requires the intermediate SKImage) is the only path for high-quality resampling.

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• dnceng.pkgs.visualstudio.com
  • Triggering command: /home/REDACTED/work/machinelearning/machinelearning/.dotnet/dotnet dotnet build src/Microsoft.ML.ImageAnalytics/Microsoft.ML.ImageAnalytics.csproj --no-restore (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

[danmoseley]

/azp run

[azure-pipelines]

Commenter does not have sufficient privileges for PR 7582 in repo dotnet/machinelearning

[danmoseley]

@stephentoub if you let me push to this I can poke on it.

[stephentoub]

Go for it

[rokonec]

CI failures are all pre-existing/flaky and unrelated to the SkiaSharp changes:

• MacOS timeouts: Known infra issue (dnceng#1883), also red on \main\
• CodeCoverage: Known repo issue (#7417), intermittent instrumentation failure
• NetFx48 TensorFlow test crash: Flaky test with 1.64% failure rate, unrelated to SkiaSharp

A CI re-run should clear the non-macOS failures. /cc @danmoseley @stephentoub

[rokonec]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 2 pipeline(s).

[codecov]

Codecov Report

❌ Patch coverage is 77.77778% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 69.55%. Comparing base (70d7603) to head (00d95ab).
⚠️ Report is 5 commits behind head on main.

Files with missing lines	Patch %	Lines
src/Microsoft.ML.ImageAnalytics/MLImage.cs	77.77%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #7582      +/-   ##
==========================================
+ Coverage   69.07%   69.55%   +0.47%     
==========================================
  Files        1483     1484       +1     
  Lines      274513   273211    -1302     
  Branches    28285    27920     -365     
==========================================
+ Hits       189625   190029     +404     
+ Misses      77503    75824    -1679     
+ Partials     7385     7358      -27     

Flag	Coverage Δ
Debug	69.55% <77.77%> (+0.47%)	⬆️
production	63.82% <77.77%> (+0.48%)	⬆️
test	89.59% <ø> (+0.07%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/Microsoft.ML.ImageAnalytics/MLImage.cs	80.64% <77.77%> (-0.66%)	⬇️

... and 19 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.