About Me

My name is Anđelo Martinović, although outside of Croatia I often become Angelo or (sadly) Andelo. I'm a computer vision researcher, originally from Zadar, a city on Croatia's Dalmatian coast. I was a research associate of the VISICS group at KU Leuven where I did my PhD under the supervision of prof. Luc Van Gool. Currently I work as the Head of Segmentation R&D at uniqFEED, a spin-off company of ETH Zürich.

Short Background

I obtained my B.Sc. title in 2008 at the Faculty of Electrical Engineering and Computing , University of Zagreb, Croatia. In my thesis, "Recognition of Number Plate Characters", I presented an approach for detection and segmentation of vehicle registration plates. In 2010 I received my M.Sc. title, with highest honors, from the same university. In my Master's Thesis, "Implementation of Computer Vision Algorithms for Embedded Systems", I designed a system for real-time video processing on a custom FPGA development board. In 2015 I defended my PhD thesis, "Inverse Procedural Modeling of Buildings", at KU Leuven, Belgium.

Publications

Doctoral Thesis: Inverse Procedural Modeling of Buildings

Doctoral Thesis: Inverse Procedural Modeling of Buildings

Andelo Martinovic
Defended on September 14th 2015.
Size: 83 MB
[BibTex
ATLAS: A Three Layered Approach to Facade Parsing

ATLAS: A Three Layered Approach to Facade Parsing

Markus Mathias*, Andelo Martinovic*, Luc Van Gool
IJCV 2016
* Denotes equal contribution.
[BibTex]
3D All The Way: Semantic Segmentation of Urban Scenes from Start to End in 3D

3D All The Way: Semantic Segmentation of Urban Scenes from Start to End in 3D

Andelo Martinovic*, Jan Knopp*, Hayko Riemenschneider, Luc Van Gool
CVPR 2015 in Boston, Massachusetts, USA
* Denotes equal contribution.
[BibTex]  [Extended abstract]  [Code
Hierarchical Co-Segmentation of Building Facades

Hierarchical Co-Segmentation of Building Facades

Andelo Martinovic, Luc Van Gool
3DV 2014 in Tokyo, Japan
[BibTex]  [Supplementary]  [Poster
Towards Semantic City Models

Towards Semantic City Models

Luc Van Gool, Andelo Martinovic, Markus Mathias
Photogrammetric Week 2013 in Stuttgart, Germany
[BibTex
Bayesian Grammar Learning for Inverse Procedural Modeling

Bayesian Grammar Learning for Inverse Procedural Modeling

Andelo Martinovic, Luc Van Gool
CVPR 2013 in Portland, Oregon, USA
[BibTex]  [Supplementary]  [Poster]  [Slides]  [Code]
Earley Parsing for 2D Stochastic Context Free Grammars

Earley Parsing for 2D Stochastic Context Free Grammars

Andelo Martinovic, Luc Van Gool
KU Leuven Technical Report
[BibTex
Hierarchical Co-Segmentation of Building Facades

A Three-Layered Approach to Facade Parsing

Andelo Martinovic, Markus Mathias, Julien Weissenberg, Luc Van Gool
ECCV 2012 in Firenze, Italy
Oral presentation, 3.6% acceptance rate
[BibTex]  [Supplementary]  [Talk]  [Slides]  [Errata]  [Updated ECP dataset annotations]  [Code]
Procedural 3D Building Reconstruction using Shape Grammars and Detectors

Procedural 3D Building Reconstruction using Shape Grammars and Detectors

Markus Mathias, Andelo Martinovic, Julien Weissenberg, Luc Van Gool
3DIMPVT 2011 in Hangzhou, China
[BibTex
Automatic Architectural Style Recognition

Automatic Architectural Style Recognition

Markus Mathias, Andelo Martinovic, Julien Weissenberg, Simon Haegler, Luc Van Gool
3D-ARCH 2011 in Trento, Italy
[BibTex
Real-time Detection and Recognition of Traffic Signs

Real-time Detection and Recognition of Traffic Signs

Andelo Martinovic, Goran Glavas, Matko Juribasic, Davor Sutic, Zoran Kalafatic
MIPRO 2010 in Opatija, Croatia
Implementation of Computer Vision Algorithms for Embedded Systems

Master's thesis (In Croatian): Implementation of Computer Vision Algorithms for Embedded Systems

Supervisor: Zoran Kalafatic, Ph.D., Faculty of Electrical Engineering and Computing
July 2010, University of Zagreb, Croatia


Datasets

Automatic Architectural Style Recognition

Updated ECP Dataset Annotations

Used in the "A Three-Layered Approach to Facade Parsing" paper.
Size: 140 kB
Automatic Architectural Style Recognition

KU Leuven Facade Dataset

Used in the "Automatic Architectural Style Recognition" paper.
Size: 457 MB