Gülcan Can
Contact
Information
E-mail: gulcan.can@idiap.ch
Website: http://www.idiap.ch/ gcan/
Personal
Information
Place and Date of Birth: Ankara(Turkey), 1988
Nationality: Turkish
Education
PhD, Electrical Engineering, Swiss Federal Institute of Technology in Lausanne, Switzerland
• September 2013 - current.
• Supervisor: Prof. Dr. Daniel Gatica-Perez
• Co-supervisor: Prof. Dr. Jean Marc Odobez
• Project: Multimedia Analysis and Access for Documentation and Decipherment of
Maya Epigraphy (MAAYA)
• Field: Computer Vision, Machine Learning, Multimedia Analysis
M.Sc., Computer Engineering, Middle East Technical University, Ankara, Turkey
• October 2010 - September 2013.
• Thesis topic: Contextual Modeling of Remote Sensing Images with Conditional Random Fields
• Supervisor: Prof. Dr. Fatoş T. Yarman Vural
• Field: Computer Vision, Pattern Recognition, Remote Sensing
• Cumulative GPA: 3.57 over 4.00
B.S., Computer Engineering, Bilkent University, Ankara, Turkey
• September 2006 - June 2010.
• B.S. Project: Question answering system
• Course Project: Object recognition using SIFT keypoints and ”bag-of-words” model
• Cumulative GPA: 3.21 over 4.00
High School, Ankara Science High School, September 2002 - June 2006.
Experience
Idiap Research Institute, Martigny, Switzerland
Research Assistant
September 2013 - current
Project: Multimedia Analysis and Access for Documentation and Decipherment of
Maya Epigraphy (MAAYA)
• Working on developing computer vision and machine learning algorithms in a
multi-disciplinary cultural heritage project which engages several archeology and
computer science partners. There are two main focus in my current research:
crowdsourcing and representation learning for glyph shapes. Crowdsourcing part
is about evaluating non-experts’ perception of Maya glyphs. For glyph representation, hand-designed shape representations and learnt representations (by
sparse auto-encoders) are compared. For further work, the emphasis will be on
representation learning.
Department of Computer Engineering, METU, Ankara, Turkey
Research Assistant
August 2010 - March 2013
Project: Processing of Remote Sensing Imagery (Land Use/Land Cover Classification and Analysis)
1 of 4
• Performing active research in a computer vision/pattern recognition project which
is run in close collaboration with several development teams in METU. The
main work consists of developing state-of-the-art algorithms to be used in classification/detection of various types of regions/objects in satellite images. For
that purpose, lots of hands-on experience is gained regarding various approaches
in feature selection (texture features based on filter responses or gray-level cooccurrence matrix, histogram-based features such as HOG and LBP, keypoint
features such as SIFT), clustering (k-means, mean-shift), segmentation (watershed, region-merging, mean-shift, graph-based), and classifier selection (SVM and
kernel selection, decision tree, k-NN, Markov and conditional random fields).
Department of Computer Engineering, METU, Ankara, Turkey
Teaching Assistant
September 2011 - September 2013
• Managing homeworks and labs for department courses such as data structures,
introduction to programming, C programming, and providing consultation to
senior design project groups.
Current and
Further
Research
Current and future research focus is on representation learning by sparse autoencoders
and deep learning architectures such as convolutional neural networks and applications
on shape classification. Previously, worked on probabilistic graphical models and their
application on visual analysis of satellite images. Have obtained a statement of accomplishment from the online course ”Probabilistic Graphical Models” taught by Prof. Dr.
Daphne Koller on Coursera.
For further study, introducing structure and other attributes into deep architectures
would be interesting and challenging.
Publications
• Gülcan Can, Jean-Marc Odobez, Daniel Gatica-Perez, Evaluating shape representations for maya glyph classification, ACM Journal on Computing and Cultural
Heritage (JOCCH), 2016. (Under review)
• Gülcan Can, Jean-Marc Odobez, Carlos Pallán Gayol, Daniel Gatica-Perez, Ancient maya writings as high-dimensional data: a visualization approach, In Digital
Humanities, 2016. (accepted)
• Rui Hu, Gülcan Can, Carlos Pallán Gayol, Guido Krempel, Jacob Spotak, Gabrielle
Vail, Stephan Marchand-Maillet, Jean-Marc Odobez, Daniel Gatica-Perez, Multimedia Analysis and Access of Ancient Maya Epigraphy: Tools to support scholars on
Maya hieroglyphics, Signal Processing Magazine, IEEE, 32(4), pp.75-84, 2015.
• Gülcan Can, Jean-Marc Odobez, Daniel Gatica-Perez, Is That a Jaguar? Segmenting Ancient Maya Glyphs via Crowdsourcing, Proc. ACM Int. Workshop on Crowdsourcing for Multimedia, 2014.
• Orhan Firat, Gülcan Can, Fatoş T. Yarman Vural, Representation Learning for
Contextual Object and Region Detection in Remote Sensing, Int. Conference on
Pattern Recognition (ICPR), 2014.
• Gülcan Can, Fatoş T. Yarman Vural, Contextual Modeling of Remote Sensing Images
with Conditional Random Fields, M.Sc. Thesis, 2013.
• Gülcan Can, Orhan Firat, Fatoş T. Yarman Vural, Contextual Object Recognition
with Conditional Random Fields, 21th Signal Processing and Communications Applications Conference, 2013.
• Gülcan Can, Orhan Firat, Fatoş T. Yarman Vural, Conditional Random Fields for
Land Use/Land Cover Classification and Complex Region Detection, 14th IAPR
International Workshop on Structural and Syntactic Pattern Recognition (SSPR,
jointly organized by ICPR), 2012.
2 of 4
• Ümit Ruşen Aktaş, Gülcan Can, Fatoş T. Yarman Vural, Edge Aware Segmentation
in Satellite Imagery: A Case Study of Shoreline Detection, 7th IAPR Workshop on
Pattern Recognition in Remote Sensing (PRRS, , jointly organized by ICPR), 2012.
• Ümit Ruşen Aktaş, Gülcan Can, Fatoş T. Yarman Vural, A Robust Approach for
Shoreline Detection in Satellite Imagery, 20th Signal Processing and Communications Applications Conference, April 2012.
• Ulya Bayram, Gülcan Can, Şebnem Düzgün, Neşe Yalabik, Evaluation of Textural
Features for Multispectral Images, SPIE Remote Sensing Conference, 2011.
• Ulya Bayram, Gülcan Can, Bariş Yüksel, Şebnem Düzgün, Neşe Yalabik, Unsupervised land use - land cover classification for multispectral images, 19th Signal
Processing and Communications Applications Conference, 2011.
Awards and
Honors
• Best paper award at 3rd Computer Science Student Workshop (2012).
• Full-time scholarship by Bilkent University based on the ranking of ÖSS 2006 (Student Selection Exam), among 2 million candidates.
• Encouragement award in biology at 14th MEF Projects Competition in high school
(2005)
Technical
Background
Programming: C++, C, Java, MATLAB, Python, HTML, Assembly
Technologies & Applications:
• Google App Engine: Coding back ends for an enterprise Java project.
• LATEX, Microsoft Office, Open Office, and similar packages.
Operating Systems: Microsoft Windows 7/XP, various Linux distributions.
Databases: MySQL.
Other: UML, XML etc.
Interests
Reading fantastic novels
Squash, swimming, skiing
Languages
Turkish (Native)
English (Fluent, Toefl Score: 99)
German (Beginner)
Japanese (Beginner)
French (Beginner)
Key Skills
Motivated self-learner, analytic problem solver
Deep understanding of computer vision and pattern recognition algorithms, with longterm implementation experience in MATLAB
Extensive development experience in object oriented languages (C++, Java) and C
References
• Dr. Daniel Gatica-Perez: Professor at Department of Electrical Engineering (EDEE),
EPFL and senior researcher at Idiap Research Institute. PhD thesis supervisor.
e-mail: gatica@idiap.ch
• Prof. Dr. Fatoş T. Yarman Vural: Professor at Department of Computer Engineering, METU. MSc. thesis supervisor.
e-mail: vural@ceng.metu.edu.tr
• Asst. Prof. Dr. Sinan Kalkan: Professor at Department of Computer Engineering, METU. Taken computer vision course and having active discussions in image
processing lab meetings weekly. e-mail: skalkan@ceng.metu.edu.tr
3 of 4
• Assoc. Prof. Dr. Selim Aksoy: Professor at Department of Computer Engineering, Bilkent University. Taken undergraduate computer vision course. e-mail: saksoy@cs.bilkent.edu.tr
4 of 4