1997-98
NEW MEXICO
HIGH SCHOOL
SUPERCOMPUTING
CHALLENGE

Interim Report


Team Number: 071
School Name: LAS CRUCES HIGH SCHOOL
Area of Science: Cryptography
Project Title: Data Embedding
Project Abstract: http://mode.lanl.k12.nm.us/97.98/abstracts/071.html
Interim Report: http://mode.lanl.k12.nm.us/97.98/interims/071.html
Final Report: http://mode.lanl.k12.nm.us/97.98/finalreports/071/finalreport.html

'A picture is worth a thousand words' as the old saying goes. This idea has held more weight than what it was meant to, especially since the development of modern data embedding. This field of technology has advanced the world of cryptography from being solely text-based to being graphically oriented. The purpose of data embedding is to completely hide the existence of data into a multimedia context. Because of its broad usage, it has been silently used to relate sensitive messages in the past and will continue to do so in the future as it is now in the present.
Through our research we have discovered that data embedding has had a rich history. Some of the first methods included shaving a messenger's head and tattooing the message on his scalp. After his hair grew back, he was sent to deliver the message. Later methods involved writing with invisible inks or compressing data onto a microfiche and writing it as a dot in a book. Today's methods are based on changing insignificant bits of the host's computer binary data. The altered files seemed to be identical to the original, yet they contained hidden information. Although modern methods of data embedding are strong, they have some weaknesses. Most of the graphic-oriented methods we have found are sensitive to distortion and/or filtration.
To embed data into an image file it is desirable not to noticably alter that image in the process. By limiting the size of the data or the amount of data embedded, we can avoid noticably contanimating the image. On the other hand, many current data embedding software packages do not provide an end product which is resistant to tampering. For example, most packages use very primitive algorithms to encrypt data; therefore, they are susceptible to degradation via distortion, filtering, cropping, resizing, truncating, etc. Our research clearly indicates the need for a technique capable of embedding data that will be strongly resistant to tampering and degradation. Our team will pursue a solution to this problem. Some solutions that we have looked into are creating a program that will use stronger methods of encrypting, using Fourier's analysis to reduce noise distortion, and finding a way to possibly combine a couple of the sofware packages that are on the market and internet.
The kind of program that we are looking to create will be able to hide a message in a picture and make it so that the message can't be destroyed but can be removed with the proper equipment. To prevent against possible multiple attacks to degrade the message, it will be encrypted to avoid trial detection. We plan on exploring several different algorithms with some based on possibly strengthening current algorithms. Some will explore entirely new tactics such as Fourier's analysis, which is similar to the use of radio. The file will represent the magnitude of the electrical field with time, and the watermark will be stored on some special frequency like in radio waves. Our other attempt at solving our problem will be in the form of combining two or more different programs. This can be done using the source codes of the programs and putting them together into such a way that the final program will be stronger than the original programs.
So, all in all, our team has researched the history of data embedding and wants to make the techniques stronger. We will attempt to solve this problem using our own programs, Fourier's analysis, and a combination of existing programs.

Team Members:

Sponsoring Teacher(s):

Project Advisor(s):


New Mexico High School Supercomputing Challenge
http://mode.lanl.k12.nm.us