[WinML] Return wgi::BitmapBounds to caller of bind

[aungthetnaing]

Is your feature request related to a problem? Please describe. When VideoFrame is "bind" to model input, it might be cropped and scaled per while maintaining aspect ratio. reference: https://github.com/microsoft/onnxruntime/blob/4e5bc8365b882563088542f877d4a83f4c08a958/winml/lib/Api/ImageFeatureValue.cpp#L126. For use case such as Object detection, the BitmapBounds info is important to map the output tensor values to detected boxes in image under evaluation. So, the collection of the BitmapBounds need to exposed to the called of the bind.

System information

• ONNX Runtime version (you are using): Microsoft.AI.MachineLearning 1.8.1

Describe the solution you'd like API need to be extended to return the BitmapBounds to the caller.

Describe alternatives you've considered

Additional context Add any other context or screenshots about the feature request here.

[stale[bot]]

This issue has been automatically marked as stale due to inactivity and will be closed in 7 days if no further activity occurs. If further support is needed, please provide an update and/or more details.

[nums11]

Thanks for posting, we'll look into this.