Yesterday Google released a visual search engine called Google Goggles. It is available right now as an application for Android 1.6+ (i.e. Donut or Eclair) based mobile devices. It uses several sophisticated image recognition techniques like object recognition, OCR, face matching, etc.
This will prove useful for m-learning – users can take a snapshot of any object and Google Goggles will provide relevant information about it. Leveraged appropriately, Goggles has great potential to be a powerful augmented reality application.
For e.g. when a user sees some landmark and wishes to get more information about it. All he need do is take a snapshot and input the image to Google Goggles which will recognize the landmark and retrieve information about it. Currently it works with book covers, landmarks, artwork, logos, places, and contact information. It doesn’t work with things like foods, cars, animals or plants.
Take a look here: