Closed ffaghri1 closed 10 years ago
Neat! The second paper even mentions inferring distance to the camera which could be useful in scenarios where we don't know the arrangement of the mics to the camera.
I briefly mentioned it in the proposal as final stages. Certainly something to look into.
I'm gonna add these papers to the proposal with a little more expanding on the subject.
@ramili and @mvcisback I have not gone through the recent proposal and issues yet, which I will do tonight. Meanwhile, I wanted to let you know of discussion I had with Minje.
@ramili and I were discussing whether we could find a connection between eigendace+eigenvoice for separation and whether the incorporation has some applications. Minje sent me these pointers:
This one seems similar to what you’re saying: http://www.cims.nyu.edu/~bregler/icassp94_bregler_konig.pdf
This one is a more general scenario: http://www.cs.technion.ac.il/~elad/publications/conferences/2005/25_CVPR_Localization.pdf