Work around CORS to fetch, process and display any image from the internet

[AmadeusW]

If we could go around CORS, then we can process any crossview image at runtime and display it in VR.
Currently, scraped images are downloaded, cropped and uploaded to CDN that sets appropriate CORS headers.

The following scenario doesn't work:

1. Load image onto canvas; HTML5 is fine
2. Copy pixels onto new canvas; CORS data is also copied
3. Crop and resize; optionally
4. Display in VR; CORS data from point 2 prevents WebVR from displaying it