Interview with assyriologist Hendrik Hameeuw, pt 2

Imagine you’re an assyriologist. You’ve investigated a clay tablet and found out that “King Darius slaughtered 20 sheep”. You want to publish that information in a journal. You can provide background information, but your paper will never get published if you can’t proof your statement. Assyriologists have basically 3 ways to give that proof.

The first one is taking plain photographs, like this example.
The problem with these photographs is that assyriologists aren’t, despite sometimes trained, good photographers. What’s even worse, lots of times the lighting conditions ‘in the field’ are far from ideal. A situation that happens more often than you would think is the following one: You get a dig permit for 1 month on the field. When do you discover most clay tablets? The day before your permit expires.
You might say that’s not so bad, just take them with you and take your time to photograph them while back home. But since the 1950ies, you cannot take them out of the country any more. They are mostly moved to a local archive or museum, and when you’re lucky, you can photograph them there. The tablets that are photographed the best are mostly the ones that are stored in European musea. A typical good photograph looks like this.

Besides photographing, a second way to prove what’s been written on a clay tablet is drawing it. Major disadvantages are that it’s a slow process and that drawings are only as good as the skills of the drawer. If he missed an important detail, you’re out of luck.

The third way of proving is: On clay tablet number 18966, stored on that shelf in that museum is written that King Darius has killed 20 sheep. Please feel free to travel to that local museum when in doubt about my statement.

But luckily there is a fourth way. The Minidome is portable. It has a uniform lighting system. It has very accurate 3D surface information. And, important for its users, it doesn’t require much user interaction, so you can do other things in the mean time. Instead of plain 2D photos, users can share and view special files containing the necessary 3D info to produce following results: (please view the 720p quality version)

Please note that the screen recording was far from ideal: the frame rate was rather poor and the compression rather high. To get an idea of the image quality, view the following image at 100%.

clay tablet, viewed with Minidome webviewer

clay tablet, viewed with Minidome webviewer

Interview with Hendrick Hameeuw

Last week, I had an interview with assyriologist Hendrik Hameeuw of the Syro-Mesopotamian Studies unit of the arts departement of the KU Leuven. I had a meeting with him to show my website and to ask him whether there are still some key features missing. Besides the technical part, we also talked about assyriology.

As I’m blogging on the preservation of cultural heritage, I think it’s really interesting to share the thoughts of someone who is actually doing this kind of work.

Cuneiform script is the oldest writing system in the world (even a bit older than the hieroglyphic script) It saw its birth in what’s nowadays southern Iraq, but what assyriologists call Mesopotamia (the land of the rivers Tigris and Eufraat) some 4000 years ago. The peoples that used this writing system were successful and got to rule over what’s now called the Middle East.

For 3000 years, cuneiform writing was used in that region. Mr Hameeuw likes to compare it with our writing system. The shapes of the characters you’ll find depend greatly on the language, region and timeperiod. As for our writing system, Gotic writing is clearly not what we are mainly using today. If you’re writing in Spanish, you’re also going to be using different characters than when writing in dutch. BTW, when writing a text in LaTeX, you might want to specify the language you’re writing in. Checkout for more info on this.

Despite all the differences, there’s luckily also some form of uniform evolution. It started as a rather difficult system with around 1500 and then somewhat later 600 signs, where each token had 1 specific meaning. The oldest texts are mostly contracts and recipients (I gave you xx for these slaves…)

After a while these ancient peoples wanted to use cuneiform writing to pass down their hymns, but of course there weren’t any symbols they could use. So they used existing symbols that sounded the same.

So, how do you get to know what’s been written on these clay tablets?
180 years ago, the western world became more interested in ancient, middle eastern relics. Mostly large statues, but they soon started wondering what all these markings on these relics were. The first good realization was that those markings were texts. Then, they started making assumptions. They were for example pretty certain that King Darius or Xerxes gave the order to make a specific piece of art (They knew of the existence of these kings, because they had fought against the ancient Greeks, whose battles are pretty well documented and passed on.) Another assumption was that it was common for a King to have a small statement at the beginning or end of a text describing who they were. Piece by piece, they started to unravel more and more, but it wasn’t until they found tablets with Aramaic summaries that they could really start deciphering, for Aramaic is the ancestor of Arabic and Hebrew alphabets.

Nowadays, some 95% of cuneiform writing is understood, but we still don’t know much about the languages that they spoke. There are 2 main languages, Akkadian and Summeric. The Akkadian language, with the Assyrian and Babylonian dialect is best understood, because of its resemblance with Hebrew and Arabic.

If you’re interested in these fascinating discoveries, you might want to start your search with

In part 2, I asked Hendrik Hameeuw how assyriologists are able to do research and why the Minidome technology could be the next standard.

Fake or Fortune

Last week, I met with a dutch art collector who was interested in the Minidome technology. He had brought 2 pieces of art with him, a painting and an etched printing plate to test under the Minidome. (
He was immediately entousiastic by the quality of the images that the Minidome can produce. He told me that in his experience, it’s really difficult to photograph a painting as accurately and with the same excellent lighting as the Minidome. However, since the Minidome is made principally for cuneiform tablets, most paintings wouldn’t fit in the dome.
But then he saw the pencil strokes in 3D and was stunned by the level of detail. He started talking about a BBC television miniseries called Fake or Fortune, in which the provenance (the history of ownership) of artworks is examined. The first episode of the first season is about a real or fake painting by the impressionist Claude Monet. If you first want to see the episode, don’t read along.

Warning: Spoiler alert!

Bords de la Seine à Argenteuil by Claude Monet, taken from

Right, so in the first episode the painting “Bords de la Seine à Argenteuil” by Claude Monet is examined.
It’s a story about an old English art collector, his not so very supportive wife, the french customs, lots of experts and new scientific technologies to examine paintings and a very powerful family that holds the authority on when it comes to works of impressionists and post-impressionists.
Proving that an artwork is painted by the old masters Rubens or Rembrandt and not by one of their pupils is a difficult task. For Rembrandt as well as Rubens there are special committees, however, for impressionists like Monet there isn’t such a group of experts that gives their final verdict. No, for Monet you’d have to rely on this catalogue:  (mind the price) by a very nowadays notorious man called Daniel Wildenstein. If it’s in this book, it’s considered to be a real Monet, if it isn’t in there, it is not.
And of course it isn’t in the book. Englishman David Joel has tried to prove for nearly 20 years that the painting he bought for 40.000 pounds is a real Monet, but with the help of the crew of Fake or Fortune, his quest has been brought to a whole new level. Infrared, X ray photography, a 240 megapixel camera examining the painting in 13 frequency bands and several experts cast their light on the subject.  Verdict: It is most likely a genuine Monet!

New imaging technologies will hopefully help experts in proving some (famous) paintings once thought authentic, fake and vica versa. In a business where lots of money can be won or lost, this is a hard task to do.

Please do not link to any youtube video of this first episode of Fake or Fortune, as that might not be entirely legal.

Digitizing our (European) cultural heritage

The main purpose of the Minidome, the 3D scanner I’m writing a viewer/analyzer for, is the digitisation of artifacts.
Digitizing our cultural heritage is important. The two main reasons for this gigantic task are access and preservation. By scanning an old book that’s hidden somewhere in a library and making it available online, multiple scolars or amateurs can view and study the book at the same time, from their own offices and homes. What’s more, by applying special techniques during scanning, you can reveal much more about the object than there is for the naked eye to see, as I explained in last post. Scanning old books once and making the copies available also reduces the wear of paging through the book.
Digitizing libraries and musea is an enormous task. One that needs to be repeated every x years because our scanning technology and understanding of the characteristics of the object to be scanned gets better and better. You can for example photograph a painting with a camera and uniform lighting. But you could easily use a 3D scanner such as the Minidome and get much more (3D) information. Questions such as how do you store metadata (period/year in which the object was made, the location it was found, the dimensions, links to papers that are written about the object…) And how do you store them digitally? Should there be one gigantic database for every artifact found anywhere in the world? That’s probably not possible. But if that doesn’t exist, how do you know where and how to search in the different online collections?
Because of the difficulty of the tasks involved, the engineering, computer science and arts faculties of universities are working together to aid musea and libraries (eg
The European Commission also pays lots of attention to this important task, as can be seen on
In a 2012 report (mind the typo in the title on the first page) by, a project funded by the European Commission to get statistical data about digitization, digital preservation and online access to cultural heritage in Europe, I’ve found some interesting numbers. Around 1/5 of all collections that needs to be, are digitized. The vast difference between art museums (42%) and national libraries (only 4%) is also striking.
So if you’re wondering what Europe’s doing with your tax money, supporting these kinds of projects is a small part of its budget (

Using a 3D scanner to scan 2D books

Yesterday I had a talk with a researcher who has helped on the development of the Minidome. I asked him why a group of scientists wants to use the Minidome to digitalize old books and manuscripts. You see, the Minidome was developped to aid in the digitalisation of clay tablets. Writing in cuneiform (spijkerschrift) is really difficult to capture. The only thing they could do was taking pictures under varying lighting conditions and drawing what they saw. The 3D shape of the writing can easily be captured by the Minidome. Thanks to some smart digital filters, the software can also produce sketches, which can be used in papers etc.

But why would you want to use the Minidome to capture books and manuscripts? Firstly, the Minidome is a good camera with an excellent, uniform lighting system. The lighting conditions are always the same. But mostly because of its ability to capture 3D information. Yes, there is 3D information on 2D pages. The pages may appear 2D, but the resolution of the Minidome is good enough to provide plenty of information invisible to our naked eyes. How hard was pressed on the feather that was used to write and draw all those symbols. Or, when in doubt about the authenticity of a painting for example, to see whether the dynamics of the paintbrush strokes correspond with those of a painting that was determined to be genuine.
When studying artifacts, you want to have as much information to analyze as possible. Even on 2D appearing surfaces like sheets of a book or a manuscript, there is 3D info.

Real time visualisation and processing of Minidome recordings

Before I’ll continue blogging about the social relevance of my thesis, I shall talk about my thesis itself.
The preliminary title is Real time visualisation and processing of Minidome recordings.
A Minidome is a 3d scanner, developped at VISICS < PSI < ESAT < KULeuven.




Schematic representation of Minidome

Schematic representation of Minidome

It consists of a dome with a diameter of 0.7m on which 264 LEDs are strategically placed. At the top, a camera is mounted.
Each LED is turned on seperately and a photo is being made. Because of the geometry and local color of the object, some parts will be brighter than others. This is repeated for every LED, so that there are 264 photos of the object. Those pictures are being processed by a special algorithm that renders a couple of textures that can be used by an OpenGL (or in my case WebGL) program to give an impression of 3D. Because there is only one camera recording from the top, only the side of the object facing the camera is being recorded. If you want to record another side, you’ll have to rotate the object and repeat the procedure.

A wireframe representation of an artifact scanned with the Minidome

A wireframe representation of an artifact scanned with the Minidome


A very very alpha version of the viewer: The yellow squares represent the direction of 2 virtual light sources.

A very very alpha version of the viewer: The yellow squares represent the direction of 2 virtual light sources.

I’m writing a platform indepedent viewer and analyser for those recordings. It’s a website, that can be viewed with Firefox, Chrome, Safari and Opera on Windows/Mac OSX/Linux computers. By slightly rotating the object and moving a virtual light it should give a good impression of the 3D shape. But next to being a viewer, it should also aid the user in examining the object, by means of some smart, yet intuitive filters.
The main users are archeologists, but it’s not farfetched to think that once a 3D model is made, a museum will show it on a computer, next to the real object, to aid visitors in viewing and apprehending the artifact.

on hard disk drive failure

You could have deleted your personal files accidentally . Your laptop might have been stolen. Or your hard disk drive might have failed. As I wrote in last post, there are various reasons why you should have backed up your data. This time I’m going to write a bit about hard disk drive failure.

I’m not going to talk about the reasons why hard disk drives fail, but about the chances that a HDD fails in the future.

A study done by Google in 2007 on more than hundred thousand HDDs reveals some interesting and contra-intuitive stuff.

For starters, every HDD has a Self-Monitoring, Analysis and Reporting Technology (SMART) test system available. You can read out SMART information with the help of some software ( The results should give an indication of the health of a HDD.
Sadly, you cannot easily check the SMART status of an external HDD since most enclosures don’t pass the relevant information to the operating system.
But what’s even more sad is that researchers that conducted the 2007 study found that although some SMART parameters are highly correlated with failure probabilities (e.g. Drives that have at least one scan error are 39 times more likely to fail in the next 2 months than drives with no scan errors), SMART data alone will not be a good predicator of drive failure, certainly not when you want to know the chances of failure of a specific drive (and not of a population)
In short, checking the SMART status alone is not enough to know whether a HDD is healthy or will fail tomorrow.

A second remarkable finding is that there is little correlation between how hard a HDD is stressed (the workload) and failure rate. Another remarkable finding is that there is little or no correlation between average drive temperature and failure rate. This is an interesting result because lots of efford and money is spent on cooling. Google lowers it’s energy bill drastically by allowing their datacenters to become a bit hotter than previously thought was safe. See also

Despite we rely on them all the time, we simply cannot predict with the tools of today when a hard disk drive will fail. The best way to make sure your data is kept safe is to use multiple backups or a proper RAID system