Amazon Remembers with experimental iPhone app

Inevitably, Amazon.com has joined the iPhone Apps Store, and while it does more or less what you'd expect, it's also a chance for Amazon to experiment with image-based search and recommendations.

 

In addition to allowing iPhone users to browse through its catalog of books, CDs, DVDs and products offered by third-party retailers, the Amazon app also includes a feature called Amazon Remembers.

 

Here's how it works (according to the NYT's Bits Blog): Take a picture of any product you see anywhere, and Amazon Remembers will upload it to the site, after which freelance workers in Amazon's Mechanical Turk program will match the product in the photo with products for sale on Amazon.com - the idea being that Amazon can offer you a better deal on the same or similar item, and thus convince you to buy it online rather than in meatspace.

 

Amazon says the feature is "experimental", and it's not hard to see why - the company fully admits that results can take anywhere from five minutes to 24 hours, and is probably more likely to work with window shoppers than people with more immediate needs.

 

Still, the idea of getting people to use camera phones as a search tool is enticing, although certainly not new. NTT DoCoMo introduced such a feature in its N902iS handset using technology from D2 Communications and Bandai Networks last year. In May this year, AT&T Wireless and 23half started trialing a service that uses MMS as a search tool.

 

What makes Amazon Remembers interesting is that it's backed by "˜crowdsourcing' - which means humans evaluate the images instead of algorithms. It may not be as cost-effective - although as I understand it, Mechanical Turk pays peanuts - but it's potentially more accurate.

Suggested Articles

Wireless operators can provide 5G services with spectrum bands both above and below 6 GHz—but that doesn't mean that all countries will let them.

Here are the stories we’re tracking today.

The 5G Mobile Network Architecture research project will implement two 5G use cases in real-world test beds.