This year the Nerd Herd Team (Riaan Zoetmulder, Emiel Hoogeboom, Wolf Vos and me) are to represent the University of Amsterdam on the international arena in Paris at the final stage of the Data Science Game 2016 competition following the successful appearance of the team represented our university in the previous year.
During the qualification round over 100 teams from 28 countries have been defeated to make us ranked in the third place with our deep net solution for the roof direction classification task. Consequently, we have been invited to participate in the final phase in Paris during 2 days in September to compete against TOP20 teams from over the world including Stanford University, Columbia University, National University of Singapore and University of Cambridge.
We would like to say many thanks to the 3DUniversum company, our main gold sponsor during the Data Science Game Competition.
3DUniversum (https://www.3DUniversum.com) provides low-cost software solutions for real-time 3D mobile scanning and object recognition. 3DU consists of a team of AI specialists with expertise in 3D computer vision and deep learning. The mission of 3DU is to become the global technology leader in advanced 3D scanning, analysis and printing technologies for B2B and B2C markets.
Paris, the Nerd Herd is coming!
Maastricht is one of the oldest city in the Netherlands. And because of its position
near the border of three countries (France, Belgium and Germany) it was supposed to
be one of the most interesting place to visit in the Netherlands in general. It was one
of the reasons why we decided to spend one of our last summer weekend at the beginning
of August to travel to it instead of Antwerp that is situated in Belgium.
Many parts of the old city walls can still be found. There were at least 3 layers of
protection, but only the southern parts of the second wall have been well preserved.
Parts of the first and third walls are hidden between old buildings.
The summer is coming to an end (Oh nooo!) together with Google Summer of Code. The time
to wrap everything up and write a bit more about what have been done in the Labplot
application where I have contributed for 3 summer months.
The goal of my GSoC project was to integrate VTK library to Labplot to make
manipulation and visualization of 3D objects possible. It should be very important part
in Labplot in the future because of the fact that it positions itself as an open-source
alternative to such kind of proprietary scientific graphing and data analysis softwares
like OriginLab where these features are well supported.
This year I had a long trip to A Coruna to participate in the Akademy conference. It was
an amazing opportunity to meet and to be among people that share the same vision and goal
to create and share free sofware to the world.
It was also my first time when I travelled such a long distance by bus (from Amsterdam to
A Coruna) that spent about 1.5 days in one way, and I hope it’s my first and last experience
with that :)
Unfortunately, this report will be quite short, because after coming back to Amsterdam I got
ill after a nice partying on the boat during the Amsterdam Gay Pride that was the immediate
continuation of the trip to Spain :D
Here you can find my photo report made during the Akademy 2015 conference. In the case
if you like to be anonymized, please notify me so I can remove you from the album. All photos
are distributed under the CC-BY-NC-SA license.
The midterm evaluation has been passed and there is the time to report about features and improvements that have been done during the last weeks of the Google Summer of Code programme.
3D visualization in LabPlot is supporting now several types of data:
- File sources (currently only .obj and .stl files)
Via spreadsheets an user can define coordinates of points and triangles that link them to visualize on a plot. Via matrixes a surface can be defined as a set of points with X and Y coordinates determined by matrix element indexes and Z coordinate values from matrix cells.