Saturday, July 4, 2009

Replace Magnetic Lock On Shower Door

LSB Method (Part 2) LSB method

In brief, the LSB method is characterized by the following:
  1. The information is hidden in the least significant bits.
  2. The method is robust, since if the watermarked image is distorted information is lost. An analysis
  3. visually LSB plane, one can see that "something strange" (visual assault), which can cause an attacker to distort the image to remove the information.
As the third point, when you mark an image with the LSB method can be noticed at a glance the distortion generated by modifying the least significant bits as shown in the previous entry. In that entry newspaper hid a message in reality it is very unlikely to have it. What we will do then is to hide text in the image. Each letter of the text will be converted to ASCII value (8 bits), so the message will hide a series of 8-bit numbers. The message is the famous hide poetry of Mario Benedetti teacher (who recently passed away) called " not save yourself." Then write an extract from the poem:

Never
salves Mario Benedetti


not stay still on the roadside does not freeze the joy

not want
reluctantly
not save yourself now or ever

not save yourself ...

The image marked with the poetry and the LSB plane is shown below:

Lena marked with a poem by Benedetti

As you can see a pattern of vertical lines due to the ASCII encoding we use, since many letters share certain bits, like the lowercase letters that belong to a small range of the ASCII encoding. Poetry hidden in the image consists of 762 Bytes. If poetry hide 5 times the image can better see the pattern of vertical lines generated:

vertical pattern generated at the LSB to hide text

To avoid this pattern generated by hiding the text in the image can hide a random mark at the LSB. Before proceeding with this amendment they will list the main changes you can make the LSB method.

The main changes are LSB Method: Hide
  1. random information in the LSB plane to avoid a visual assault.
  2. information
  3. Hide image edges to avoid homogeneous zones to avoid visual attacks. Using more levels
  4. LSB for hide information. Use
  5. LSB method to detect changes made in the image. Using the method
  6. LSB as a reversible method to extract the original image of the brand.
The first modification is to randomly hide bits of the message in the image. The path we will take the brand random mess. The way to do it after taking the bitstream of the mark will proceed to scramble it using a key. This key will be necessary to rearrange the mark after removal of the marked image. Then observe the marked image and the LSB plane of the same. To do the same poetry hide Benedetti described above.

Lena marked with a random message

We can realize a visual attack have no effect if the mark randomly hide because we can not take in some pattern or something that looks strange in LSB plane. The following image shows the significant difference of not using and using a random mark.

Comparison of LSB planes marked using original and random marks

Another way to make this amendment is to randomly distribute the brand in the entire LSB plane.

The first amendment prevents a visual attack, but there a type of attack called statistical chi-square that can detect the presence of the mark in the marked image as long as the mark is hidden in a random .

In the next post will describe the Chi square test, then we will continue with the other four remaining LSB modification method.

Greetings to all.

PD. The source code (in Matlab) of the methods discussed in this post are listed below:
1. LSB Method hiding a message in an image bmp paper (described in the previous post).
2. LSB Method hiding text in a bmp image.
3. LSB Method hiding text in a bmp image randomly.

0 comments:

Post a Comment