Geo tagging for better speech recognition

Here’s an interesting way to improve voice recognition performance……

Geotagged Voice Recognition

Voice recognition on mobile devices is harder than for desktop machines because of the prevalence of background noise. Now U.S. patent 8175872 demonstrates two things about this: First it shows Google expects voice recognition systems to become pretty ubiquitous, and it suggests a way to use geotagging to remove background noise and thus improve recognition.

The idea’s pretty simple: If there are several folk speaking to their mobile devices in an area, chances are they’ll all be sampling some of the same background sounds. By geotagging your voice uplink, a cloud server somewhere could use that information to identify background noises and then subtract those signals from everyone’s audio feed. Brilliant. Clever. And quite definitely creepy.

Project Glass

We know Google wants to change the world with a revolutionary augmented reality device called Project Glass, and we know what it looks like because we’ve seen a few in the wild already. Now Google’s moved to patent the design of Glass’s headset in three different versions: One is the high-tech “headband” version that’s been shown already, another is just the frame for that headband without a prominent over-eye projection module, and the final is for a fashion-conscious version that looks more like traditional sunglasses.

Great to see Google’s concerned that we all don’t look like extras from Star Trek when wearing Glass, but what we really want to see is more data on how the actual things work. Guess we’ll have to wait.

Chat about this news with Kit Eaton on Twitter and Fast Company too.

Fast Company by Kit Eaton

More-latest speech technologies
Social share or comment – what do you think?

There are no comments yet. Be the first and leave a response!

Leave a Reply

Wanting to leave an <em>phasis on your comment?

 
Trackback URL http://www.speechtechnologygroup.com/geo-tagging-for-better-speech-recognition/trackback

Geo tagging for better speech recognition

Here’s an interesting way to improve voice recognition performance……

Geotagged Voice Recognition

Voice recognition on mobile devices is harder than for desktop machines because of the prevalence of background noise. Now U.S. patent 8175872 demonstrates two things about this: First it shows Google expects voice recognition systems to become pretty ubiquitous, and it suggests a way to use geotagging to remove background noise and thus improve recognition.

The idea’s pretty simple: If there are several folk speaking to their mobile devices in an area, chances are they’ll all be sampling some of the same background sounds. By geotagging your voice uplink, a cloud server somewhere could use that information to identify background noises and then subtract those signals from everyone’s audio feed. Brilliant. Clever. And quite definitely creepy.

Project Glass

We know Google wants to change the world with a revolutionary augmented reality device called Project Glass, and we know what it looks like because we’ve seen a few in the wild already. Now Google’s moved to patent the design of Glass’s headset in three different versions: One is the high-tech “headband” version that’s been shown already, another is just the frame for that headband without a prominent over-eye projection module, and the final is for a fashion-conscious version that looks more like traditional sunglasses.

Great to see Google’s concerned that we all don’t look like extras from Star Trek when wearing Glass, but what we really want to see is more data on how the actual things work. Guess we’ll have to wait.

Chat about this news with Kit Eaton on Twitter and Fast Company too.

Fast Company by Kit Eaton

More-latest speech technologies
Social share or comment – what do you think?

There are no comments yet. Be the first and leave a response!

Leave a Reply

Wanting to leave an <em>phasis on your comment?

 
Trackback URL http://www.speechtechnologygroup.com/geo-tagging-for-better-speech-recognition/trackback
css.php