Computer Vision Engineer

3/2014 – offen

I am working on an independent project which involves the development of a labeling and annotation application. The application is using image processing and machine learning techniques for aiding the user in the process of manual labeling.

Maschinelles Lernen, mySQL, SCRUM, Agile Entwicklung, Software Design, UML, Enterprise Architect (EA), Apache Subversion (SVN), J2SE (Java SE), NetBeans, Jira

12/2013 – 7/2016

During the first year I was part of a team that developed and validated a traffic signs recognition system based on image processing and machine learning. The project was a production one.

My first responsibility was to develop the tool chain for the training of the classifiers used in the project. Having the tool ready, I trained the classifiers, conducted missed scene analysis, checked positives and negatives to make sure they are good for training, improved the training time, provided better results etc. The proposed tool was able to train 300 independent classifiers in parallel.
Another task I had was related to the study of using the colour channel for the detection of the traffic signs (used as pre-processing or as focus of attention).

In the second year I was involved in the forward looking lane recognition project. The project was already available, therefore my first responsibility was to port it to the current Adasens framework and understand the implementation, find weak points and propose solutions. Having the forward looking lane recognition system running on different types of data, with very good performance, my second big task was to apply the same system for rear view cameras (mounted in the mirrors). Taking into account the difference of perspective and all the implications, a big amount of work had to be done and it still going on. The project is a prototype.

In order to fulfill my tasks I used SVN, Jira, Visual Studio, Netbeans, Adasens framework for development and data analysis.

Maschinelles Lernen, Software Design, UML, Enterprise Architect (EA), Apache Subversion (SVN), C, C++, J2SE (Java SE)

2/2012 – 5/2013

I was part of a team that adapted and tested the software for a radar sensor (Continental radar ARS 300), based on customer specifications. My main responsibilities were to developed and maintain the diagnosis module.

Beside development, I was the Software Integrator for the project and I was taking care of quality related tasks for the project. I was conducting integration tests, organising code review meetings, wrote requirements, checked the documentation provided by the testing team etc.

In order to fulfill my tasks I used MKS, Source Insight, Visual Studio, Lauterbach, Doors, Rhapsody, CAN protocol applications (PEAK PCAN, CANalyzer, CANoe).

Project type: production project.

Software Analyse, C, Softwareentwicklung (allg.), Visual Studio