> MMI’s primary implementation is through the use of a JavaScript tag which interrogates the device to which the ad was delivered for the presence of a graphical processing unit (GPU). The JavaScript creates a Canvas Element in HTML5, which allows access to WebGL (Web Graphics Library), a JavaScript API for rendering interactive graphics within any compatible web browser without the use of plug-ins. An IVT classification decision is made based on the results of the rendering capability of the device.

How this can’t be emulated by a headless browser?