Tag Everything

As mentioned in the previous post, given any visual information (image, video), it’s still difficult for computers to recognize the object that contained in the information. Although a lot of works have been done in this area, we still have only limited progress. Unfortunately, in the near future, we still can’t be too optimistic about the progress towards the field. But the good news is that we do have an easier solution: tag everything!

An example mentioned by Prof. Ramesh Raskar in the MIT Media Lab is supermarkets. In a supermarket we can recognize all the merchandise not because we have a super advance AI system there. We can recognize the objects because we’ve used a workaround: barcode. What if we can have tags for every object in our world? Just as the following video shows (beginning around 2:08), we can let the device we have to interact with them. This is especially convenient as we living in an era of smart phones.

Imagine you just entering a supermarket, and your smartphone app automatically shows you the flyer and highlight the items that are recommended for you based on your taste, shopping history, etc. What’s more, you can browse the highlighted object in an augmented reality fashion, such as the following video.

Of course, there are far more applications other than merchandise in supermarkets or bookstores, etc. For instance, for someone you would like to know, you can take a picture of he or she (or “sense” the tag, depending on the implementation) using your smart phone, then the Facebook page of that person will prompt out on your browser (of course we should empower users to have control over their privacy).

The key point for the above mentioned vision is to tag everything, before computers can recognize objects by AI or CV algorithm. It’s highly possible given the advance of tagging technology such as RFID or Bokode. Probably in the near future we can enjoy the convenience by tagging everything.

---