Tuesday, June 12, 2007

Please Welcome Asirra

Asirra Project
In some sites, we have seen a lot of random words (or characters) that has to be inputted when we want to submit something (register to a forum, add a comment to a weblog, register an email account, etc). It's called CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) or HIP (Human Interactive Proof). The purpose of this technology is clear: to tell whether the submitter is human or computer (in forms of bots, which might be a spammer in most cases). In most cases, bots are unable to see the characters since it's being displayed as an image, so they have to analyze it deeper in order to solve that puzzle (even though there has been a lot of reports that nowadays spammers has find a way to read an image text created by CAPTCHA).

The other problem with CAPCTHA is that sometimes, the words is too scrambled or fuzzed, so it has created more problems than solving problems. Some people had to retry for several times before they can be considered pass the test. For something important, this is acceptable, but if for something simple, this is annoying.

For that reason, a new project has been started by Microsoft, called Asirra (Animal Species Image Recognition for Restricting Access). Instead of using words, Asirra will ask the audience to identify photos of cats and dogs. This task difficult for computers, but people can accomplish it quickly and accurately (unless they haven't seen a dog or a cat or any pet before). Asirra will use Petfinder as their partner for providing them with the images they need. So far, they provide two millions of images for Asirra project, which is quite huge (remember, we can randomly mix this images, so there are a lot of combinations here). Asirra's idea comes from HOTCAPCTHA which uses "hot" human photos to proof that the audience is human or computer.

Right now, the current status is still in beta, but if you are interested, you can try it, since Microsoft has provide a detailed information about how to use it in their installation page. It uses Javascript for the client's components and Microsoft provides the web services which will be called from the client. The returned value will be in XML format. Microsoft has provide some examples for some programming language, including Python, PHP (and a PHP WordPress blog plugin), C#, VB, JScript, and Perl (but you can add it by modifying the source with the language you use in your daily work).