Fighting the WebBots

1
Fighting the WebBots

• A webbot is a program that visits web sites for
  all kinds of purposes.
• For example, Google webbots make copies of all
  web sites for their search engines.
• The challenge is to stop malicious webbots

2
Webbots and Spam

• Spammers send webbots to get e-mail accounts from
  sites that offer them for free.
• How can you tell that someone who asks for an
  e-mail account is a person or a webbot?

3
Are you a person or a bot?

• We know that there are certain things that
  computers cannot do.
• Ask the applicant to do something that
  computers cannot do.
• Cook a meal?
• Read something impossible for computers to read!

4
CAPTCHA

• CompletelyAutomatedPublicTuring test to
  tellComputers andHumansApart

5
CAPTCHA

• CAPTCHA does not have to be text, but computer
  unreadable text is convenient.
• Alternatives include pictures.
• For example, ask if a person in a pictures is
  smiling or not. What is wrong with such a CAPTCHA

6
How Computers ReadOptical Character Recognition
(OCR)

• Step 1 Separate print (usually dark) from
  background (usually light).
• Step 2 Pick up individual characters (group of
  dark pixels)
• Step 3 Identify their shape by looking for
  strokes, loops, corners, etc
• Step 4 Use rules to classify. For example, an H
  has two vertical strokes and a short horizontal
  strokes.

7
Frustrating OCR
8
Make your own CAPTCHA

• A web site that offers you the means
• www.codeproject.com/KB/aspnet/CaptchaImage.aspx
• For a general tutorial see
• www.theopavlidis.com/technology/captcha/tutorial.
  htm

9
Some Weak CAPTCHAs
From Paypal
From YahoosBriefcase
10
Some CAPTCHAs that may betoo hard for people
From Yahoo
From Passport
11
But Human Vision is Amazing
12
But Human Vision is Amazing
13
Non Text CAPTCHAs

• Use pictures as CAPTCHAs
• Plus There are very tough to break
• Minus
• Need to label a huge number of pictures.
• If we use few pictures the webbot can just keep
  guessing.

14
Synthetic Pictures(an idea by M. Kaplan)
Please click on or enter each letter
corresponding to the following list in the field
below. You must enter them in the exact sequence
listed.
C
K