Google researchers have published a paper (PDF) which was awarded as best paper in Euro Interactive Television Conference last week. The research says, with an audio clip recorded using simple computer microphone we can find relative web-content. They used this technique to demonstrate a simple audio clip from running conversion can find relevant web content. The ambient noise doesn’t hamper the search quality and accuracy. Amazing! I think, we are not very far when we could implement a voice / video & image search. There are billions of phones which have capability to capture images, record sounds and video clips. Simplest use can be capture a photo of a person to whom you don’t know and send it to online search engines using web-services. Then the services will find the images which contain the person and you will have all the information about the person in seconds (provided he has public existence on web). Though this sounds like movie scene from Matrix or Mission Impossible but it’s very useful for guys like me who forgets people faces. (I have many incidents of talking to people to whom I don’t remember anymore and just pretend that I know them ;))
Another application is very important for me about audio search. I heard a music snippet from an online movie or some place and I don’t know what kind of music that is or any artist information. Now, how should I use search engines to find the song name, artist name etc. The online audio search may be very useful in this sense, send an audio clip and you will receive list of songs which contain that audio portion. Amazing… I am so much excited to see all these technologies in action.
Although there are couple of issues regarding media protection, online privacy and other things to be solved but compared to facilities gained by end user, those issues doesn’t seem big.