221–230 of 238 posts
Earlier quoted context omitted.
Here is an example of a problem that would be hard for computer to solve: http://www.youtube.com/watch?v=4G4y79ZbaBs
You just diff each frame and keep the parts that don't change much. Very simple to solve.
This is an unsolved computer vision problem.
Earlier quoted context omitted.
I'd be curious to see the ability for a machine to solve picture based captcha systems. For example, given a lineup of 10 pictures of pets, choose the three that are cats. I've seen them before, just not widely implemented.
Even just randomly picking 3 images will be successful about 1% of the time which is more than enough for a bot.
A forum I run in Ireland doesn't have a need for Captchas. The forum is very regional and niche - no genuine foreign visitors - so we blocked every country apart from Ireland and the UK from posting comments. Result - no spam. This wouldn't work for many forums, but if it's local, you really don't need to open it to the world.
Earlier quoted context omitted.
This is computer vision, an I'm somewhat an expert in the area. I can tell that video sequence recognition is _much_ harder problem than image recognition. For example, if you would show an letter made out of random noise moving through random noise, current computer vision algorithms would not be able to recognize anything. And you would pick out that letter immediately. Human visual subsystem is really amazing in t…
It should be possible to do this with an animated GIF. Do you have any references/examples I could use as a starting point?
Should be relatively easy to code with any library that can draw a text on a bitmap. Like PIL, matplotlib, etc. Use ffmpeg to make a video out of frames.
1. draw letters (just black/white) masks; 2. fill letters with noise; 4. fill background with noise; 5. copy letters using a mask onto background, using X,Y as loc; 6. add a little bit of new noise to letters; 8. modify X,Y coordinates (move letters SLIGHTLY); 9. go to step 4.
Every time I read something like this I pine for never having the opportunity to turn MotionCAPTCHA[0] into a real (secure) system. Everyone said it couldn't be done, but I know it can! [0] http://www.josscrowcroft.com/demos/motioncaptcha/
The only "protection" it offers is that, as long as MotionCAPTCHA does not become very popular, it won't be worth the effort for a bot spammer to write the circumventing code.
However, you can get exactly the same amount of protection without inconveniencing your users by obfuscating field names and doing the same JavaScript form action replacing trick you already do.
The simple problem is this: The challenge to properly submitting the form is tied to the fact whether your plugin has replaced the form action url to contain the right value. But this is not tied in any way to the user actually having solved the motion captcha. Which means that any bot-stopping power it possesses can also be achieved equally well without forcing the user to pass the motioncaptcha test.
And while the motion captcha looks really neat and fun, it is still always a better user experience if they can submit a form without having to jump through a hoop at all.
People interested in capchas may also like this episode of the Hacker Medley podcast http://hackermedley.org/humans-only/ It talks about capchas vs ocr technologies and capchas solving companies in India.
Is that no longer true? If so, I should look at reviving our old system.