Visual Captchas

[migcosta]

Hey guys,
First of all please don't consider this an act of spam, it is not my intencion to spam!

I belong to a small team that created the visualCaptcha (visualcaptcha.net), and we are looking for some reviews from web developers.
We have downloads but we don't get much feedback :/ so I decided to come here and ask you guys what do you think about this.

So really what I would like to know is:
what id your opinion on Visual Captchas?
do you think they are safe?
what would make you try it?

Many thanks
cheers

[IIM]

• Even if someone drags the image a bit away from your drop box,it is validating it.
• Second thing ,it can be modified as position can be modified by bot and he can modify and check as there are only 5 images and it is fixed so there is a probability of 0.2 that user selects correct image.
• Also if the images and matching words are fixed user can check that also.
• Third point the name(like cat,world) should be image(that too with some noise effect so as not easily character recognisible by bot) not character otherwise anyone can view source and then find image accordingly.

[migcosta]

hello IIM tks for your feedback :)
1 - the position of the dragged icon: you can move the icons around but the captcha will only consider the one in the "drop here area" :)
The validation is done server side not by js. So the captcha is validated only after the form submission. Every time the page is loaded you will have a new visualCaptcha. The only way to know if a visualcaptcha is valid is by submitting the form.. not by dropping random images :)

3 - to do an action you need to click, drag and drop. The captcha wont work just by changing the position of the element. You will need a visual bot to do it: you need ocr or an html reading library; you need a visual recognition tool; and you will need to have a single machine just for using that bot.

4 - visualCaptcha is completely configurable, you can set your own images and texts (and name your images as you wish) :) but your concern has given me an interesting idea to make this even more difficult to crack :D

5 - "Third point the name" nice idea :)

once again tks for you feedback it was great and helped us a lot :D
cheers

[IIM]

@migcosta -
1) it is considering box as drag box instead of circle as shown in image.
2) I am talking about bot ,for them crawling image and placing at drop box is easy and success rate it too high approx 0.2.Make the position of image also random.As for now 5 images location are fixed.Try to give a bit random position so as it is hard for bot.
3) Try to do things from backend.i.e. using JQuery/AJAX to get the 5 images so that it is not visible in source page.