Run semantic segmentation sample on full GPU pipeline

[huningxin]

This is required by webmachinelearning/webnn#226

To implement this pipeline, the semantic segmentation sample probably could execute following steps:

1. Use mediacapture-transform to capture the VideoFrame
2. Import the VideoFrame into a WebGPU texture (GPUExternalTexture)
3. Execute WebGPU shader to do pre-processing and convert the GPUExternalTexture to a GPUBuffer
4. Compute the MLGraph with converted GPUBuffer and the output is another GPUBuffer
5. Execute WebGPU shader to do post-processing on the output GPUBuffer and render the result to a canvas.
6. Create a VideoFrame from the CanvasImageSource and enqueue to the mediacapture-transform API's controller.

[huningxin]

Related to webmachinelearning/webnn-polyfill#156

[huningxin]

depends on w3c/webcodecs#412