Fault convert to ONNX object detection model

**System Information:**
- OS & Version: Windows 10
- ML.NET Version: ML.NET v3.0.0-preview.23266.6
- .NET Version: .NET 7.0

**Describe the bug**
Fault conversion to ONNX

**To Reproduce**
Steps to reproduce the behavior:
1. Train object detection model by tutorial (https://devblogs.microsoft.com/dotnet/object-detection-ml-dotnet-model-builder/)
2. Try convert model
```charp
public class ModelInput
{
[LoadColumn(0)]
[ColumnName(@"Labels")]
public string[] Labels { get; set; }

[LoadColumn(1)]
[ColumnName(@"Image")]
[Microsoft.ML.Transforms.Image.ImageType(640, 640)]
public MLImage Image { get; set; }

[LoadColumn(2)]
[ColumnName(@"Box")]
public float[] Box { get; set; }

}

public static void Export()
{
var onnxPath = "test.onnx";
var mlContext = new MLContext();
ITransformer mlModel = mlContext.Model.Load(MLNetModelPath, out var _);

var input = new ModelInput[30];
for (int i = 0; i < input.Length; i++)
{
input[i] = new ModelInput
{
Image = MLImage.CreateFromFile(@"demo.jpg")
};
}
IDataView trainingData = mlContext.Data.LoadFromEnumerable(input);
using (var onnx = File.Open(onnxPath, FileMode.OpenOrCreate))
{
mlContext.Model.ConvertToOnnx(mlModel, trainingData, onnx);
}
}
```
3. Got error
System.InvalidOperationException: 'Unsupported type: Microsoft.ML.Data.MLImage'

**Expected behavior**
Got onnx model

**Additional context**
If I try to represent the input as an array [B,3,W,H], it says that it expects an image, when I pass the image, it says that it does not support it.