Re: new scenario suggestion: aquire media capture without user interaction

On 2012-07-06 07:51, Sunyang (Eric) wrote:
> *Scenario*: Face detection without user interaction with camera.
>
> *Description*:
>
> A authentication device using a browser is set on a entrance gate of
> meeting room, the device has camera and microphone, and no
> keyboard/mouse/touch screen etc. The browser is always connected with a
> control center web system. One man come to the meeting room, and make
> face near the device, the js app start capture video from camera without
> request user permission, then make recognition with the captured video.
>
> *Requirement derived:*
>
> We should let getUserMedia has a way to tell the UA we do not need user
> permission to capture video, and the UA need to have some configuration
> switch about it.

I recognize that we will have requirements where installed (thus 
trusted) apps and white-listed apps may directly access the camera. 
However, I think we could derive those from a more common use case. With 
that I mean when the ua is obviously a regular browser with all the 
regular input methods.

In this particular case, I would guess that the person who starts up the 
face recognition software would give the app permission to use the 
camera and then the app would have that permission for the whole 
lifetime. Since the camera is the only input it must be active all the 
time to know when a face is near and it should start processing.

/Adam

Received on Friday, 6 July 2012 08:56:17 UTC