Use ARCamera as input for MediaPipe?

[patrick508]

Hi,

So for a project I'm working on I got the mediapipe sample to work with the webcam as done in the Sample project as well. However in my project I use the ARCameraManager from Unity to render the camera image to the screen. I need this camera because I am also trying to get the depth from this camera.

This is currently giving me issues as MediaPipe tries to start and access the webcam, while ARCamera is already using the camera. I tried to make the MediaPipe sample to work with the ARCamera but failed. It's tightly coupled to the webcam as for now. Is there any input or help I could get regarding this issue? Perhaps someone has already managed to get it to work with the ARFoundation ARCameraManager?

In short what I'm trying to achieve: Give MediaPipe the texture2D I get from ARCameraManager(I managed to get this texture already) and get the pose from that source.

[homuler]

In the following, I assume that you are using PoseTrackingSolution.

It's tightly coupled to the webcam as for now.

Not really.
Currently, texture data is read as follows.

MediaPipeUnityPlugin/Assets/Mediapipe/Samples/Scenes/Pose Tracking/PoseTrackingSolution.cs

Lines 134 to 135 in 2994e44

	// Copy current image to TextureFrame
	ReadFromImageSource(imageSource, textureFrame);

MediaPipeUnityPlugin/Assets/Mediapipe/Samples/Common/Scripts/Solution.cs

Lines 88 to 108 in 2994e44

	protected static void ReadFromImageSource(ImageSource imageSource, TextureFrame textureFrame)
	{
	var sourceTexture = imageSource.GetCurrentTexture();
	// For some reason, when the image is coiped on GPU, latency tends to be high.
	// So even when OpenGL ES is available, use CPU to copy images.
	var textureType = sourceTexture.GetType();
	if (textureType == typeof(WebCamTexture))
	{
	textureFrame.ReadTextureFromOnCPU((WebCamTexture)sourceTexture);
	}
	else if (textureType == typeof(Texture2D))
	{
	textureFrame.ReadTextureFromOnCPU((Texture2D)sourceTexture);
	}
	else
	{
	textureFrame.ReadTextureFromOnCPU(sourceTexture);
	}
	}

So if you have Texture2D at hand, you can read it like this.

// ReadFromImageSource(imageSource, textureFrame); 

// Texture2D texture2d;
textureFrame.ReadTextureFromOnCPU(texture2d);

If you want to do it the right way, you will have to implement an ImageSource class that supports ARCamera.
See ImageSource for more details.

[ROBYER1]

I would also suggest trying out the XRCPUImage with AR Foundation, let me know if you get any further with this
https://docs.unity3d.com/Packages/com.unity.xr.arsubsystems@4.1/api/UnityEngine.XR.ARSubsystems.XRCpuImage.html

[patrick508]

@homuler Thanks for the help! Sorry for the late reaction. I managed to find a solution that is kind of hacky but it works for now. Planning on cleaning it up soon and trying to find a more low level fix.

For those interested:
`
public class ARPoseProvider: BasePoseProvider, IPoseProvider
{
public IEnumerator Start()
{
AssetLoader.Provide(new StreamingAssetsResourceManager());

        StartCoroutine(StartARCameraImagePoseTracking());
        
        yield return StartCoroutine(InitializeInferenceMode());
        
        initialized = true;
    }
    
    private IEnumerator StartARCameraImagePoseTracking()
    {
        yield return StartCoroutine(InitializeARCameraSources());
    
        yield return StartCoroutine(InitializeARCameraPoseGraphRoutine());
        
        arCameraManager.frameReceived += OnARCameraFrameReceived;
    }
    
    private void OnARCameraFrameReceived(ARCameraFrameEventArgs eventArgs)
    {
        arTextureToUpdate = GetCurrentColorTexture();

        UpdateARPose(arTextureToUpdate);
    }
    
    private void UpdateARPose(Texture2D texture)
    {
        arTextureFrame ??= new TextureFrame(texture.width, texture.height);
        
        arTextureFrame.SetPixels32(texture.GetPixels32());

        arCameraPoseTrackingGraph.AddTextureFrameToInputStream(arTextureFrame).AssertOk();
        currentARCameraPose = arCameraPoseTrackingGraph.FetchNextValue();
    }

    unsafe Texture2D GetCurrentColorTexture()
    {
        if (!arCameraManager.TryAcquireLatestCpuImage(out XRCpuImage image))
            return null;

        XRCpuImage.ConversionParams conversionParams = new XRCpuImage.ConversionParams
        {
            inputRect = new RectInt(0, 0, image.width, image.height),
            outputDimensions = new Vector2Int(image.width, image.height),
            outputFormat = TextureFormat.RGBA32,
            transformation = XRCpuImage.Transformation.MirrorX
        };

        int size = image.GetConvertedDataSize(conversionParams);
        NativeArray<byte> buffer = new NativeArray<byte>(size, Allocator.Temp);

        image.Convert(conversionParams, new IntPtr(buffer.GetUnsafePtr()), buffer.Length);
        image.Dispose();

        if (arTexture == null)
        {
            arTexture = new Texture2D(conversionParams.outputDimensions.x, conversionParams.outputDimensions.y, conversionParams.outputFormat, false);
        }

        arTexture.LoadRawTextureData(buffer);
        arTexture.Apply();
        buffer.Dispose();

        return arTexture;
    }
}`

[homuler]

I also have written a minimal code to run the Face Detection solution for those who've gotten to this issue.
Please also refer to it.

// Copyright (c) 2021 homuler
//
// Use of this source code is governed by an MIT-style
// license that can be found in the LICENSE file or at
// https://opensource.org/licenses/MIT.

using Mediapipe;
using Mediapipe.Unity;

using System;
using System.Collections;

using Unity.Collections;
using Unity.Collections.LowLevel.Unsafe;
using UnityEngine;
using UnityEngine.XR.ARFoundation;
using UnityEngine.XR.ARSubsystems;

using Stopwatch = System.Diagnostics.Stopwatch;

public class ARCameraManagerTest : MonoBehaviour
{
  [SerializeField] private ARCameraManager _cameraManager;
  [SerializeField] private TextAsset _configText; // attach `face_detection_gpu.txt`

  private CalculatorGraph _calculatorGraph;
  private NativeArray<byte> _buffer;
  private Stopwatch _stopwatch;
  private ResourceManager _resourceManager;
  private GpuResources _gpuResources;

  private IEnumerator Start()
  {
    _cameraManager.frameReceived += OnCameraFrameReceived;
    _stopwatch = new Stopwatch();

    _resourceManager = new StreamingAssetsResourceManager();
    yield return _resourceManager.PrepareAssetAsync("face_detection_short_range.bytes");
    yield return _resourceManager.PrepareAssetAsync("face_detection_full_range_sparse.bytes");

    _gpuResources = GpuResources.Create().Value();
    _calculatorGraph = new CalculatorGraph(_configText.text);
    _calculatorGraph.SetGpuResources(_gpuResources).AssertOk();

    _calculatorGraph.ObserveOutputStream("face_detections", 0, OutputCallback, true).AssertOk();

    var sidePacket = new SidePacket();
    sidePacket.Emplace("input_rotation", new IntPacket(0));
    sidePacket.Emplace("input_horizontally_flipped", new BoolPacket(false));
    sidePacket.Emplace("input_vertically_flipped", new BoolPacket(true));
    sidePacket.Emplace("model_type", new IntPacket(0));

    _calculatorGraph.StartRun(sidePacket).AssertOk();
    _stopwatch.Start();
  }

  private void OnDestroy()
  {
    _cameraManager.frameReceived -= OnCameraFrameReceived;

    var status = _calculatorGraph.CloseAllPacketSources();
    if (!status.Ok())
    {
      Debug.Log($"Failed to close packet sources: {status}");
    }

    status = _calculatorGraph.WaitUntilDone();
    if (!status.Ok())
    {
      Debug.Log(status);
    }

    _calculatorGraph.Dispose();
    _gpuResources.Dispose();
    _buffer.Dispose();
  }

  private unsafe void OnCameraFrameReceived(ARCameraFrameEventArgs eventArgs)
  {
    if (_cameraManager.TryAcquireLatestCpuImage(out var image))
    {
      InitBuffer(image);

      var conversionParams = new XRCpuImage.ConversionParams(image, TextureFormat.RGBA32);
      var ptr = (IntPtr)NativeArrayUnsafeUtility.GetUnsafePtr(_buffer);
      image.Convert(conversionParams, ptr, _buffer.Length);
      image.Dispose();

      var imageFrame = new ImageFrame(ImageFormat.Types.Format.Srgba, image.width, image.height, 4 * image.width, _buffer);
      var currentTimestamp = _stopwatch.ElapsedTicks / (TimeSpan.TicksPerMillisecond / 1000);
      var imageFramePacket = new ImageFramePacket(imageFrame, new Timestamp(currentTimestamp));

      _calculatorGraph.AddPacketToInputStream("input_video", imageFramePacket).AssertOk();
    }
  }

  private void InitBuffer(XRCpuImage image)
  {
    var length = image.width * image.height * 4;
    if (_buffer == null || _buffer.Length != length)
    {
      _buffer = new NativeArray<byte>(length, Allocator.Persistent, NativeArrayOptions.UninitializedMemory);
    }
  }

  [AOT.MonoPInvokeCallback(typeof(CalculatorGraph.NativePacketCallback))]
  private static IntPtr OutputCallback(IntPtr graphPtr, int steramId, IntPtr packetPtr)
  {
    try
    {
      using (var packet = new DetectionVectorPacket(packetPtr, false))
      {
        var value = packet.IsEmpty() ? null : packet.Get();

        if (value != null && value.Count > 0)
        {
          foreach (var detection in value)
          {
            Debug.Log(detection);
          }
        }
      }
      return Status.Ok().mpPtr;
    }
    catch (Exception e)
    {
      return Status.FailedPrecondition(e.ToString()).mpPtr;
    }
  }
}

[pinak1999]

I also have written a minimal code to run the Face Detection solution for those who've gotten to this issue. Please also refer to it.

// Copyright (c) 2021 homuler
//
// Use of this source code is governed by an MIT-style
// license that can be found in the LICENSE file or at
// https://opensource.org/licenses/MIT.

using Mediapipe;
using Mediapipe.Unity;

using System;
using System.Collections;

using Unity.Collections;
using Unity.Collections.LowLevel.Unsafe;
using UnityEngine;
using UnityEngine.XR.ARFoundation;
using UnityEngine.XR.ARSubsystems;

using Stopwatch = System.Diagnostics.Stopwatch;

public class ARCameraManagerTest : MonoBehaviour
{
  [SerializeField] private ARCameraManager _cameraManager;
  [SerializeField] private TextAsset _configText; // attach `face_detection_gpu.txt`

  private CalculatorGraph _calculatorGraph;
  private NativeArray<byte> _buffer;
  private Stopwatch _stopwatch;
  private ResourceManager _resourceManager;
  private GpuResources _gpuResources;

  private IEnumerator Start()
  {
    _cameraManager.frameReceived += OnCameraFrameReceived;
    _stopwatch = new Stopwatch();

    _resourceManager = new StreamingAssetsResourceManager();
    yield return _resourceManager.PrepareAssetAsync("face_detection_short_range.bytes");
    yield return _resourceManager.PrepareAssetAsync("face_detection_full_range_sparse.bytes");

    _gpuResources = GpuResources.Create().Value();
    _calculatorGraph = new CalculatorGraph(_configText.text);
    _calculatorGraph.SetGpuResources(_gpuResources).AssertOk();

    _calculatorGraph.ObserveOutputStream("face_detections", 0, OutputCallback, true).AssertOk();

    var sidePacket = new SidePacket();
    sidePacket.Emplace("input_rotation", new IntPacket(0));
    sidePacket.Emplace("input_horizontally_flipped", new BoolPacket(false));
    sidePacket.Emplace("input_vertically_flipped", new BoolPacket(true));
    sidePacket.Emplace("model_type", new IntPacket(0));

    _calculatorGraph.StartRun(sidePacket).AssertOk();
    _stopwatch.Start();
  }

  private void OnDestroy()
  {
    _cameraManager.frameReceived -= OnCameraFrameReceived;

    var status = _calculatorGraph.CloseAllPacketSources();
    if (!status.Ok())
    {
      Debug.Log($"Failed to close packet sources: {status}");
    }

    status = _calculatorGraph.WaitUntilDone();
    if (!status.Ok())
    {
      Debug.Log(status);
    }

    _calculatorGraph.Dispose();
    _gpuResources.Dispose();
    _buffer.Dispose();
  }

  private unsafe void OnCameraFrameReceived(ARCameraFrameEventArgs eventArgs)
  {
    if (_cameraManager.TryAcquireLatestCpuImage(out var image))
    {
      InitBuffer(image);

      var conversionParams = new XRCpuImage.ConversionParams(image, TextureFormat.RGBA32);
      var ptr = (IntPtr)NativeArrayUnsafeUtility.GetUnsafePtr(_buffer);
      image.Convert(conversionParams, ptr, _buffer.Length);
      image.Dispose();

      var imageFrame = new ImageFrame(ImageFormat.Types.Format.Srgba, image.width, image.height, 4 * image.width, _buffer);
      var currentTimestamp = _stopwatch.ElapsedTicks / (TimeSpan.TicksPerMillisecond / 1000);
      var imageFramePacket = new ImageFramePacket(imageFrame, new Timestamp(currentTimestamp));

      _calculatorGraph.AddPacketToInputStream("input_video", imageFramePacket).AssertOk();
    }
  }

  private void InitBuffer(XRCpuImage image)
  {
    var length = image.width * image.height * 4;
    if (_buffer == null || _buffer.Length != length)
    {
      _buffer = new NativeArray<byte>(length, Allocator.Persistent, NativeArrayOptions.UninitializedMemory);
    }
  }

  [AOT.MonoPInvokeCallback(typeof(CalculatorGraph.NativePacketCallback))]
  private static IntPtr OutputCallback(IntPtr graphPtr, int steramId, IntPtr packetPtr)
  {
    try
    {
      using (var packet = new DetectionVectorPacket(packetPtr, false))
      {
        var value = packet.IsEmpty() ? null : packet.Get();

        if (value != null && value.Count > 0)
        {
          foreach (var detection in value)
          {
            Debug.Log(detection);
          }
        }
      }
      return Status.Ok().mpPtr;
    }
    catch (Exception e)
    {
      return Status.FailedPrecondition(e.ToString()).mpPtr;
    }
  }
}

@homuler Error at line 39. I'm using the latest MediaPipeUnityPlugin-all.zip
Assets\MediaPipeUnity\Samples\Scenes\Face Detection\AR-MP\ARCameraManagerTest.cs(39,64): error CS0407: 'IntPtr ARCameraManagerTest.OutputCallback(IntPtr, int, IntPtr)' has the wrong return type

[homuler]

@pinak1999 Please see #803 (comment)

[dogadogan]

I'm also getting the same error CS0407: 'IntPtr ARCameraManagerTest.OutputCallback(IntPtr, int, IntPtr)' has the wrong return type.

@homuler - How is #803 (comment) related to this issue? Because it seems like that comment is using lifted_objects, whereas the code you have on this thread is for face_detections, right?

Doesn't face_detections use a vector the way you already have it?

[homuler]

NativePacketCallback, the 3rd argument of CalculatorGraph.ObserveOutputStream, should return StatusArgs now and that's why the above code has compile errors.

MediaPipeUnityPlugin/Packages/com.github.homuler.mediapipe/Runtime/Scripts/Framework/CalculatorGraph.cs

Line 16 in 01cdd56

public delegate StatusArgs NativePacketCallback(IntPtr graphPtr, int streamId, IntPtr packetPtr);

You may rewrite the OutputCallback, but I recommend you use OutputStream instead of calling CalculatorGraph#ObserveOutputStream directly.

[dogadogan]

I see, thanks! I tried the suggested approach but I'm getting the following error related to the OutputCallback:
error CS0246: The type or namespace name 'FrameAnnotation' could not be found (are you missing a using directive or an assembly reference?)

Any thoughts? Here's the code I added:

 private IEnumerator Start()
  {
    // ...
    _faceDetectionsStream = new OutputStream<FrameAnnotationPacket, FrameAnnotation>(_calculatorGraph, "face_detections");
    _faceDetectionsStream.AddListener(OutputCallback);
    _calculatorGraph.StartRun(sidePacket).AssertOk();
    // ...
  }
private void OutputCallback(object stream, OutputEventArgs<FrameAnnotation> eventArgs)
  {
    Debug.Log(eventArgs.value);
  }

[homuler]

If you want to use FaceDetection, see the FaceDetectionGraph example.

MediaPipeUnityPlugin/Assets/MediaPipeUnity/Samples/Scenes/Face Detection/FaceDetectionGraph.cs

Lines 78 to 79 in 01cdd56

	_faceDetectionsStream = new OutputStream<DetectionVectorPacket, List<Detection>>(
	calculatorGraph, _FaceDetectionsStreamName, config.AddPacketPresenceCalculator(_FaceDetectionsStreamName), timeoutMicrosec);

Note that the output type of face_detections is not FrameAnnotation.

# Detected faces. (std::vector<Detection>)
output_stream: "face_detections"

https://github.com/google/mediapipe/blob/0dee33ccba37fcb9362a90b0042cd46730a7f9b5/mediapipe/graphs/face_detection/face_detection_desktop_live.pbtxt#L8-L9

See also #803 (comment).

Incidentally, the FrameAnnotation class is used in the Objectron sample, which is deprecated now.

[dogadogan]

Thank you! I fixed the output type and now it works well for face detection :)

I actually wanted to do this ARCore integration for object detection, but gave face detection a try as the first step.
So now I adjusted the code for object detection (i.e., by using "output_detections" etc.), and the connection seems to be working!

Now I want to make use of the object detector output by adding more lines into OutputCallback, however, even if I add something small, it gives me this error:

Error Unity MediaPipeException: INTERNAL: Graph has errors: 
Error Unity System.NullReferenceException: Object reference not set to an instance of an object.
 Error Unity   at ARMPObjectDetection.OutputCallback (System.Object stream, Mediapipe.Unity.OutputEventArgs`1[TValue] eventArgs) 
 Unity   at Mediapipe.Unity.OutputStream`2[TPacket,TValue].InvokeIfOutputStreamFound (System.IntPtr graphPtr, System.Int32 streamId, System.IntPtr packetPtr) [0x00000] in 
Error Unity   at Mediapipe.Status.AssertOk ()
Error Unity   at ARMPObjectDetection.OnCameraFrameReceived (UnityEngine.XR.ARFoundation.ARCameraFrameEventArgs eventArgs) [
Error Unity   at UnityEngine.XR.ARFoundation.ARCameraManager.InvokeFrameReceivedEvent (UnityEngine.XR.ARSubsystems.XRCameraFrame frame) [0x00000] in
Error Unity   at UnityEngine.XR.ARFoundation.ARCameraManager.Update ()

What I wanted to do is similar to #1037 (comment).
I understand that eventArgs.value is a System.Collections.Generic.List.
But even printing the number of detected objects (Debug.Log("Object count is: " + eventArgs.value.Count); in OutputCallback) gives the above error.

Do you know how I could address this issue? :)

[dogadogan]

Ah, alright, based on #935 (comment), I realized I should use a question mark, i.e., eventArgs.value?.Count, to ensure the value is not null.

[dogadogan]

Hi! I just wanted to follow up on something related to this ARcore integration :)

I realized that the object detector/classifier works much better if the phone is held at a certain orientation. For example, it works much better if I hold the phone in landscape mode, compared to portrait mode.

I was wondering if there is a way to set the orientation of MediaPipe detection somehow in the code?

[KiranJodhani]

@homuler I am trying to implement this but I am not getting it working. Below is the steps i followed

1. I created sample script which has a listener from arCameraManager. I can see it's working

Now I have Texture2D from OnARCameraFrameReceived and applying this to textureFrame.ReadTextureFromOnCPU(OutputTextureFromARCamera); (This is not in ImageSourceSolution and I am using holistic)

From here I am not sure I am doing it correctly or not
In Bootstrape

I changed from webcam to image but when i run showing texture in canvas. so for testing I added sample texure manually in static image source and it showed correctly. My question here is can we keep imagesource as image and set texture generated from OnARCameraFrameReceived? I tried this but when i change texture it's not updating somehow

Please share your thoughts

[homuler]

@dogadogan

I was wondering if there is a way to set the orientation of MediaPipe detection somehow in the code?

At least, you can rotate the input image on the Unity side

MediaPipeUnityPlugin/Assets/MediaPipeUnity/Samples/Scenes/Tasks/Face Detection/FaceDetectorRunner.cs

Line 80 in 2d2863e

var req = textureFrame.ReadTextureAsync(imageSource.GetCurrentTexture(), flipHorizontally, flipVertically);

or using ImageTransformationCalculator.

MediaPipeUnityPlugin/Assets/MediaPipeUnity/Samples/Scenes/Object Detection/object_detection_gpu.txt

Lines 65 to 78 in 2d2863e

	node: {
	calculator: "ImageTransformationCalculator"
	input_stream: "IMAGE_GPU:throttled_input_video_gpu"
	input_side_packet: "ROTATION_DEGREES:input_rotation"
	input_side_packet: "FLIP_HORIZONTALLY:input_horizontally_flipped"
	input_side_packet: "FLIP_VERTICALLY:input_vertically_flipped"
	output_stream: "IMAGE_GPU:transformed_input_video"
	node_options: {
	[type.googleapis.com/mediapipe.ImageTransformationCalculatorOptions] {
	output_width: 320
	output_height: 320
	}
	}
	}

[homuler]

@KiranJodhani Will you create a new issue? I'm sorry but I'm not sure what is the problem.

[KiranJodhani]

@homuler Thanks for the swift reply. I have created new issue here. Please feel free to ask any question you have about the issue. It's feature request though

#1045

[dogadogan]

Super helpful reply (#343 (comment)), thank you @homuler! :)

As a follow-up, I'm wondering if there is a way to access the originating input frame in an executed OutputCallback. For instance, when running Object Detection, only certain frames result in successful classification (whereas some frames are no good due to motion blur, etc., so then the returned value eventArgs.value in OutputCallback is null).

Do you know if there is a way to get the successful camera frame when OutputCallback returns non-empty eventArgs.value?
Because I would like to do further processing on the original image then, based on the results in eventArgs.value.
Or perhaps there's a smarter way to do this?

[homuler]

The Task API is designed to receive input images through a callback,

MediaPipeUnityPlugin/Packages/com.github.homuler.mediapipe/Runtime/Scripts/Tasks/Vision/FaceDetector/FaceDetectorOptions.cs

Line 27 in 2d2863e

public delegate void ResultCallback(Components.Containers.DetectionResult detectionResult, Image image, int timestampMs);

but OutputStream is not, so it may be difficult if not impossible.

If it's acceptable, you can get the same result by executing the graph synchronously. In this case, probably you have the reference to the input image after getting the result.

[dogadogan]

Oh great, good to know, thanks @homuler!

Do you have any example code on how to use Task API instead of OutputStream/OutputCallback in this repo or elsewhere?
My code is working well with the previous setup, so it's a bit hard for me to know where to start readjusting it for this.

Also, any pointers on how to execute the graph synchronously would be helpful too!
In my use case, ARCore would run continuously in the background for the main AR tasks, and as long as it's relatively fast, it should be ok I believe. But at the same time, I wonder if this would interfere with ARCore/Foundation's OnCameraFrameReceived...

[homuler]

Do you have any example code on how to use Task API instead of OutputStream/OutputCallback in this repo or elsewhere?

https://github.com/homuler/MediaPipeUnityPlugin/blob/2d2863ea740a6a5ad01854ea88ab5f48be2a36b6/Assets/MediaPipeUnity/Samples/Scenes/Tasks/Face%20Detection/FaceDetectorRunner.cs

Note that ObjectDetector is not ported yet.

Also, any pointers on how to execute the graph synchronously would be helpful too!

See https://github.com/homuler/MediaPipeUnityPlugin/wiki/Getting-Started#get-imageframe.
Alternatively, if you have ample memory, you can keep the image for reference and search for an image with the same timestamp as the output timestamp.

[dogadogan]

Great, thanks @homuler! Any idea when ObjectDetector would be ported for Task API?

Thinking about it, I feel like the synchronous approach might cause problems for real-time ARCore. Have you tried this with AR solutions in the past?

Your idea of storing the last few frames in an array for reference, and later searching for the one with the same timestamp as the output timestamp in the OutputCallback makes sense! However, do you know how I could access the originating timestamp from within OutputCallback (so I can use it for comparison across the captured frames stored in the array)?

[dogadogan]

Hi @homuler! :) I would appreciate if you could share any insights you might have about the comment above ^^
Thank you for your help and time!!

[homuler]

Any idea when ObjectDetector would be ported for Task API?

Maybe when I feel motivated. If not pressured by others, it might be achievable by around next month.

However, do you know how I could access the originating timestamp from within OutputCallback

Ah, it's designed not to pass the timestamp to the callback, so you'll need to make modifications to the following lines.

MediaPipeUnityPlugin/Packages/com.github.homuler.mediapipe/Runtime/Scripts/Unity/OutputStream.cs

Line 400 in 2d2863e

outputStream.OnReceived?.Invoke(outputStream, new OutputEventArgs<TValue>(value));

P.S. Responding to closed issues is challenging, so if necessary, please create a new issue.