Europass

Curriculum Vitae

Personal information
First name(s) / Surname(s) Gellért Sándor, Máttyus
Affiliation and address Uber ATG Toronto
661 University Ave, Suite 720
Toronto, Ontario, M5G 1M1
Canada
Telephone(s) +1 437 886 9822
E-mail gmattyus@uber.com; mattyusgege@gmail.com
Webpage http://gellertmattyus.info/

Nationality Hungarian

Gender male

Occupation
Engineer: Machine Learning, Computer Vision, Self-Driving

Work experience
Dates 2017 - Ongoing
Occupation or position held Engineer
Main activities and responsibilities Research in Computer Vision and Machine Learning related to Self-Driving. Software engineering and
development related to the research.
Name and address of employer Uber ATG
661 University Ave, Suite 720
Toronto, Ontario, M5G 1M1
Canada
Type of business or sector Research and Development department of a private company.

Dates 2016 - 2017

Occupation or position held Postdoctoral Fellow
Main activities and responsibilities Research in Computer Vision, Machine Learning and Remote Sensing.
My research focuses on creating and enhancing road maps automatically from aerial images and
ground images jointly. During my work, I investigate and propose new methods, implement it in
computer programs and publish the results in international conferences and journals.
Name and address of employer University of Toronto, Department of Computer Science.
6 King's College Road, Pratt building
Toronto, Ontario, M5S 3G4, Canada
Type of business or sector Public University

Dates 2011 - 2016

Occupation or position held Research Fellow
Page 1/4 - Curriculum vitae of For more information on Europass go to http://europass.cedefop.europa.eu
Gellert Mattyus © European Communities, 2003 2017.2.2.
 Main activities and responsibilities Research and software development in the following topics and projects: Road layout parsing and
vehicle detection on aerial images, image geolocalizaton and orthorectification without GPS and IMU
and software development for data transmission from airplane to ground station in the DLR project
VABENE (Real time monitoring of natural disasters, mass events, and large traffic disasters).
In the VABENE project I was working on a system processing aerial images real-time on the board of
an airplane to predict the traffic on the ground. This involved to work on a system where the hardware,
the software and the machine learning must work together. I was involved in this project for 5 years.
By this I have gained experience how software and hardware decisions effect the project in the later
stages.
Automatic vessel detection on optical satellite images for the European Maritime Safety Agency.
Name and address of employer German Aerospace Center (DLR), The Remote Sensing Technology Institute, Photogrammetry and
Image Analysis (PB)
Type of business or sector Public Research Organisation

Dates 2010-2011
Occupation or position held Embedded software developer
Main activities and responsibilities Development and maintenance of Ericsson Mini-Link (microwave radio link installed with embedded
Linux).
Name and address of employer Ericsson Hungary R&D
Type of business or sector Research and Development department of a private company.

Dates 2008– 2010

Occupation or position held Student researcher, developer
Main activities and responsibilities Research in image processing and computer vision. Multi target tracking on aerial videos. The
algorithms were implemented in C++ and Matlab.
Thesis: Motion trajectory recognition in video image-sequences
Name and address of employer Computer and Automation Research Institute of the
Hungarian Academy of Sciences (MTA SZTAKI), Budapest, Hungary
Type of business or sector Public Research Institute

Education and training

Dates 2004 – 2010

Title of qualification awarded Diploma in Electrical Engineering
Principal subjects/occupational skills Electrical Engineering, Embedded Systems, Signal Processing, Image Processing, Computer Vision
covered
Name and type of organisation Budapest University of Technology (BME), Budapest, Hungary
providing education and training
Level in national or international Hungarian equivalent to MSc.
classification
Grades average 4.92 (5.0 is the best)
Prizes 2005: Physics competition of the faculty: 2nd prize.

2006: „Electrical Signals and Systems” competition of the faculty: 3rd prize.

2006-2007: "Scholarship of the Hungarian Republic" for excellent school achievements.

Dates 2013 - 2016

Title of qualification PhD degree with “Summa cum Laude”.
Name and type of organisation University of Technology, Munich (TUM), Munich, Germany. Chair of Remote Sensing Technology.
providing education and training
Supervisors Professor Richard Bamler, Dr. Friedrich Fraundorfer

Page 2/4 - Curriculum vitae of For more information on Europass go to http://europass.cedefop.europa.eu

Gellert Mattyus © European Communities, 2003 2017.2.2.
 Topic Joint Information Augmentation of Road Maps, Aerial Imagery and Ground Images.

Dates 2014 July and August, 2015 February and March, 2015 August and September
Title of qualification Visiting PhD student
Name and type of organisation Department of Computer Science, University of Toronto
providing education and training
Supervisor Professor Raquel Urtasun
Topic Enhancing World Maps by Parsing Aerial and Ground Images

Self-assessment Understanding Speaking Writing

European level (*) Listening Reading Spoken interaction Spoken production
English C1 C1 C1 C1 C1
German C1 C1 C1 C1 C1
Spanish A2 A2 A2 A2 A2
(*) Common European Framework of Reference for Languages

Social skills and competences Working in national and international teams and projects. Project planning.

Organisational skills and Guide for trainees, master thesis supervisor.

competences

Technical skills and competences Basic electrical engineering skills: electrical and signal measuring, Verilog hardware description
language: design on Xilinx FPGA-s.

Computer skills and competences User skills: MS Windows, MS Office, Linux, Latex, GIT, Phabricator
Developer skills: Software development on Windows, Linux, embedded Linux and embedded operation
systems.
Computer languages: C, C++, C#, Java, Python, Matlab, Bash script, Windows batch script.
Frameworks: PyTorch, TensorFlow, Spark
Computer Network knowledge and skills.
Image processing, computer vision and machine learning knowledge and skills: object detection,
tracking, image-based localization, image registration, deep neural networks, graphical models, image
segmentation, remote sensing images, geospatial data.

Driving licence Category B

Awards and Prizes: IEEE Geoscience and Remote Sensing Society 2016 Letters Prize Paper Award (Paper of the year in
a journal).

2006: 3rd prize in the „Electrical Signals and Systems” competition of the Faculty of Electrical
Engineering at the Budapest University of Technology.

2005: 2nd prize in the “Physics” competition of the Faculty of Electrical Engineering at the Budapest
University of Technology.

Scholarships and Fellowships: 2017: “2017 Postdoctoral RBC Innovation Fellowship” (University-wide 1 person) at the University of
Toronto.

2006-2007: "Scholarship of the Hungarian Republic" for excellent school achievements.

Page 3/4 - Curriculum vitae of For more information on Europass go to http://europass.cedefop.europa.eu

Gellert Mattyus © European Communities, 2003 2017.2.2.
 Publications: Bai, Min; Mattyus, Gellert; Homayounfar, Namdar; Wang, Shenlong; Lakshmikanth, Shrinidhi and
Urtasun, Raquel: Deep Multi-Sensor Lane Detection. IEEE/RSJ International Conference on Intelligent
Robots and Systems (IROS) 2018.

Mattyus, Gellert and Urtasun, Raquel: Matching Adversarial Networks. Conference on Computer
Vision and Pattern Recognition (CVPR) 2018.

Mattyus, Gellert; Luo Wenjie and Urtasun, Raquel: DeepRoadMapper: Extracting Road Topology from
Aerial Images. IEEE International Conference on Computer Vision (ICCV) 2017.

Wang, Shenlong; Bai, Min; Mattyus, Gellert; Chu, Hang; Luo, Wenjie; Yang, Bin; Liang, Justin;
Cheverie, Joel; Fidler, Sanja and Urtasun, Raquel: TorontoCity: Seeing the World with a Million Eyes.
IEEE International Conference on Computer Vision (ICCV) 2017.

Mattyus, Gellert and Fraundorfer, Friedrich: Aerial image sequence geolocalization with road traffic as
invariant feature. Image and Vision Computing, 52 (8), pp. 218-229. 2016

Mattyus, Gellert; Wang, Shenlong; Fidler, Sanja and Urtasun, Raquel: HD Maps: Fine-grained Road
Segmentation by Parsing Ground and Aerial Images. In: Conference on Computer Vision and Pattern
Recognition (CVPR) 2016.

Mattyus, Gellert; Wang, Shenlong; Fidler, Sanja and Urtasun, Raquel: Enhancing Road Maps by
Parsing Aerial Images Around the World. In: IEEE International Conference On Computer Vision
(ICCV) 2015.

d'Angelo, Pablo; Mattyus, Gellert and Reinartz, Peter: Skybox image and video product evaluation. In
International Journal of Image and Data Fusion, vol. 6, Nov. 2015

Liu, Kang; Mattyus, Gellert: Fast Multiclass Vehicle Detection on Aerial Images. In Geoscience and
Remote Sensing Letters, IEEE, vol.12, no.9, pp.1938,1942, Sept. 2015

Mattyus, Gellert and Fraundorfer, Friedrich: Aerial image mosaicking with online calibration - A
feasibility study. In: Proceeding of Conference of the Hungarian Association for Image Processing and
Pattern Recognition (online), 2015.

Kurz, Franz and Rosenbaum, Dominik and Meynberg, Oliver and Mattyus, Gellert and Reinartz, Peter:
Performance of a real-time sensor and processing system on a helicopter. In: ISPRS Archives, pp.
189-193. ISPRS Archive. Pecora 19 Symposium in conjunction with the Joint Symposium of ISPRS
Technical Commission I and IAG Commission 4, Nov. 2014

Mattyus, Gellert: Near real-time automatic vessel detection on optical satellite images. In: ISPRS
Hannover Workshop, Volume XL-1/W1, pp. 233-237. ISPRS Archives. ISPRS Hannover Workshop
2013, 21 May - 24 May 2013, Hannover, Germany.

Mattyus, Gellert; Kurz, Franz; Rosenbaum, Dominik and Meynberg, Oliver: A real-time optical airborne
road traffic monitoring system. In: Hungarian Association for Image Processing and Pattern
Recognition, pp. 645-656. Hungarian Association for Image Processing and Pattern Recognition.
KEPAF 2013, 29 Jan - 01 Feb 2013, Bakonybel, Hungary.

Mattyus, Gellert; Benedek, Csaba and Sziranyi, Tamas: Multi target tracking on aerial videos, ISPRS
Istanbul Workshop 2010 on Modelling of optical airborne and spaceborne sensors, WG I/4, Oct. 11-
13, IAPRS Vol. XXXVIII, part 1/W4.

Toronto, Canada, 10.12.2018

Page 4/4 - Curriculum vitae of For more information on Europass go to http://europass.cedefop.europa.eu

Gellert Mattyus © European Communities, 2003 2017.2.2.