| September 2007 | ||||||||
|---|---|---|---|---|---|---|---|---|
| Sun | Mon | Tue | Wed | Thu | Fri | Sat | ||
| 1 | ||||||||
| 2 | 3 | 4 | 5 | 6 | 7 | 8 | ||
| 9 | 10 | 11 | 12 | 13 | 14 | 15 | ||
| 16 | 17 | 18 | 19 | 20 | 21 | 22 | ||
| 23 | 24 | 25 | 26 | 27 | 28 | 29 | ||
| 30 | ||||||||
Hi there,
today I read funny news I have to write about right now ;)
For several years now, forms on web pages get "secured" by 'captchas', small images with smehow scrambled text.
![]()
The intention behind that is, that machines like spam robots can't read them, but humans do.
Disturbing the background or juggling with characters should protect from ocr.
Unfortunately this 'protection' is getting weaker and weaker, because of the massive computing power of bot nets, 'social captcha decoding' (redirecting captchas to own pages, letting your own human clients decode them), and last but not least highly intelligent detection/decoding tools ... :(
Since the modern internet is somehow a new kind of "Wettrüsten" (german for 'arms race'), between the good the bad & the ugly, new strategies are in constant development ...
The latest idea: let the cats do the work (could certainly also be dogs, birds, (car) brands, etc.) :)
E. g. the KittenAuth project creates 'catchas', an other project uses a database of homeless pets/cats, so it also has the intention to find new homes for the little ones :)
Ths might be only one 'next step' in image recognition (and for sure will be broken again in near future), but to me it shows again the unbelieveable reservoir of yet unthought ideas, that's actually what i love about computer science :)
Send a message
Search for members
