Welcome back to Beyond the Patterns. So I haven't been recording videos for quite some

time now but I really wanted to go ahead and really show you some of the things that we've

been doing in the past. In particular we had really really interesting visitors from

all across the globe virtually in our lab and I would continue to highlight their presentations

in this lecture series. And one of the best presentations in the past year was actually

given by Robert Zablatnik. So he is a professor at the TU Vienna and he has actually born in

Klaugoford in Austria in 1965. Then he was from 92 to 2003 assistant professor and from 2003

to 2010 associate professor of computer vision at the pattern recognition and image processing

group at the TU Vienna. Then from 2005 and 2017 he was the head of the Institute of Computer

Aided Automation and since 2010 he is heading the computer vision lab which is part of the

2018 founded Institute of Visual Computing and Human Centro Technology which he is heading

since 2019. His research interests are 3D computer vision including rangefinder, stereo

vision, shape from X, registration, calibration, robot vision, machine and deep learning for

computer vision, video data analysis, automated document analysis and multi spectral images

as well as virtual and augmented reality and applications. And in particular he has really

awesome contributions in cultural heritage preservation and we have been meeting for a long

time in the time machine context so we have been also working in this project together

and I am really really glad that we had him here as an invited speaker and now I am able

to present his talk that he was giving here which was entitled multi spectral imaging

and writer identification for historical manuscripts.

Robot it's a great pleasure that you have been here and now the stage is yours.

Thank you for this nice introduction and of course also for giving me the possibility

to talk a little bit about things we do here at the computer vision lab and yeah today

we are going to hear something about multi spectral imaging and writer identification

I hope it works now. So maybe if you have introduction about the computer vision lab

is so we are a group of about 35 people working in computer vision as the name says and

especially in the area of three division document analysis machine and deep learning of course

and more contracting and we have also a couple of applications we are working on mainly medical

image processing and astral vision cultural heritage where we will hear more today so

the events and the assisted living. So these are the people here we have about 35 people

four of which are postdocs and four faculty assistants here, three professors and a lot

of pre-do master students and you see the blue ones here are faculty of the university

the resties on soft money that means we also have a couple of projects always running

currently it's about 20 projects some of them are sponsored by the you some of them of

the local research funding agencies and there are also some projects with the ministries

and federal easter eggs. We have about an an alternative of two million years and about

22 food and equipment and people that are produced in a couple of papers per year and approximately

10 master thesis are finished and one or two PhD thesis in a year here at this group.

Now today we're going to focus a little bit on the document analysis where we do almost

everything and is concerned with the document analysis however we start with imaging and

here multispectal imaging and maybe this is also kind of interesting to you and then I

selected the right recognition identification as one of the topics of course we do also

OCR, HDR, layout analysis and other things. Now what can you expect? No, in the first

part we will hear a little bit about multispectal imaging and why we think that multispectal

imaging is important in the area of culture heritage. How the image of a precision is

done and what can be improved because most of the people do not care about the image

acquisition in this area however for getting good results is necessary to think also about

the image acquisition especially in multispectal imaging and then we see what we can do with

the data acquired and what can be done for the ones that need this data especially for

the humanists that actually then use this information for their research and as to see