Projects:2015s1-32 Code Cracking: Who Murdered The Somerton Man?

Supervisors

• Prof Derek Abbott
• Dr Matthew Berryman

Honours students

• Nicholas Gencarelli
• Jikai Yang

Project guidelines

• Project Handbook

General project description

In this project you will attempt to solve a possible murder that took place in Adelaide in 1948. This crime remains unsolved till today, but you can use engineering to bring our knowledge closer to the killer. You can read the details about the dead body and the circumstances [1]

Associated with the dead body was this secret code:

MRGOABABD

MTBIMPANETP

MLIABOAIAQC

ITTMTSAMSTGAB

(See the original photograph, as there may be an extra line, and some of the M's may be W's. Some people also think that the last "I" is really a "V". Also the last G is probably really a C). To this day code crackers have been unable to decrypt it.

We also want you to bring the skills of an electrical engineer to bear on the area of e-forensics and see if you can apply these to other areas of the case (eg. graphical reconstruction of the dead man's face).

Specific tasks

Here are the remaining tasks resulting from previous work. You may want to focus on a subset of these:

• Critically review the statistical analysis of the letters. See if you can extend it (eg. testing another language previous students missed and checking if they included all possibilities of ambiguous letters). Is the conclusion of previous team correct? Because there are lots of interesting new tasks (see below) don't spend too long on this. Just spend enough time on this as a quick warm up exercise.

• Download the searchable pdf file of the copy of the Omar Khayyam that closely matches the dead man's copy. Create an ascii file with the raw text. Use this as a one-time pad that directly substitutes the letters of the alphabet a-z. The book contains 75 quatrains (four sentence poems) each containing about 140 letters. So the whole book contains about Failed to parse (MathML with SVG or PNG fallback (recommended for modern browsers and accessibility tools): Invalid response ("Math extension cannot connect to Restbase.") from server "https://en.wikipedia.org/api/rest_v1/":): {\displaystyle 75 \times 140 = 10,500\,} letters. As we don't know where in the book the one-time pad starts, start at the beginning and step through the whole book one letter at a time. You'll then end up with >10,000 decrypts. Write a software script to look for the most common top-20 words in all the decrypts to narrow down to a few possible results that can be examined by eye.

• Extend the CipherGUI 2011 software that was created by a previous team. See if you can add more ciphers to the collection. Use it to eliminate more ciphers and enter your conclusions here: Cipher Cross-off List. Be critical and be prepared to question and recheck some of the items already on the list.

• A previous team created a webcrawler and search engine to search keywords with wild cards, as Google does not allow this. This is to check for common repeated expressions on the WWW that may contain initial letters that are also in the code. If the code is an initialism, this will give us a clue as to some likely content. There are two things that need to be fixed: (i) we need you to write a convenient web-interface for this search engine, and (ii) we need to scrap the webcrawler as it takes too long. It is impossible to crawl the whole web with one little PC and therefore you need to interface the search engine to operate on an index that has already been created by a commercial webtrawler. Unfortunately, companies like Google won't let you access their index list. Therefore you need to use another provider such as YaCy. You may also want to check out a resource called CommonCrawl. To use CommonCrawl you'll need to sign up with both github (to download the Java source code) and also Amazon (to run your uploaded compiled code).

• Use computer graphics to reconstruct and undistort the face of the dead man. What would he look like if he were alive? To do this you need the data from the previous group that scanned the man's face from a plaster bust, at the Police Museum, with a 3D scanner. An example of the type of graphics software you can use to manipulate the scanned image is 123D. You may want to investigate other 3D rendering graphics software.

• Use the departmental 3D printer to recreate a scaled down version of the bust, before and after your 3D rendering. The motivation for creating a 3D representation is so that we can create 2D pictures of the man at any angle. It will not be long before companies like Google release next generation search engines that search for faces on the web. So having multiple images at a number of angles will be of future importance for a large-scale image search.

• Investigate the prices and availability of more expensive 3D scanners than the David 3D Scanner that we have. Can you find a 3D scanner that would have the resolution to pick up all the pores and texture on the bust of the Somerton Man?

• Plot, present, and interpret the mass spectrometer data we have.