"you'd need to create some seriously smart image processing"
Thats what I thought but I hoped there would be a nother way.
But well I guess Ill have a go at that then...lol
As soon as I have got something Ill post it here!
reCAPTCHA is intended to raise the probability that the entity on the other end of the line is a real person by making it very expensive (in developer costs, developer knowledge, and computational time) to create software that masquerades as a person. The only way for you to programatically get around reCAPTCHA is to reduce these costs by dropping the value of your developer time to near zero, do lots of research (get smart), and if you're hitting it often enough, to use powerful enough hardware that the computational cost is mitigated. In other words, you have to be rich, studious, and willing to work a long time for free. If you possess all of these qualities, I would think there are more worthwhile projects to donate your time and resources to.
This strikes me as interesting, because on a NPR show, not too long ago, I heard the inventer of Captcha say that he was working on a system, whereby users could be novelly employed as book translators. The way it worked, is if a computer scan couldn't reproduce the book's page photo as readable text, those text fragments would be forwarded to the Captcha database, and would be sent out to end-users to see how many different eyes might translate it.
That's probably referring to reCAPTCHA, or one of the newer CAPTCHA derivatives such as duolingo. The original CAPTCHA was just a means of increasing defense in the escalating arms race between bots and those who wanted to avoid them.