IVANA TOSIˇ C´ Address: Redwood Center for Theoretical Neuroscience Tel: 1 (510) 213 9579 575A Evans Hall, MC# 3198 [email protected] Berkeley, CA 94720-3198, US rctn.org/wiki/Ivana Tosic

RESEARCH INTERESTS /image processing; Computational neuroscience; ; Sparse representations; Multi-view ; Binocular vision; Omnidirectional imaging; Image and video compression; Distributed and coding; .

EDUCATION • PhD degree (May 1, 2009) 2004-2009 EPFL/IEL/LTS4, Switzerland Thesis title: On unifying sparsity and for image-based 3D scene representation Thesis advisor: Prof. Pascal Frossard

• Doctoral School in and Sciences at I&C 2003-2004 EPFL, Switzerland main courses: Information theory, and applications, Image Processing

• Electrical Engineer Diploma (Dipl.Ing. in EE) 1997-2003 Faculty of Electronic Engineering/Department of telecommunications University of Niˇs,Serbia, GPA 9.80 (max 10), best graduate award

WORK EXPERIENCE • Postdoctoral fellow November 2009 - Present Redwood Center for Theoretical Neuroscience, University of California, Berkeley, US

• Postdoctoral fellow April 2009 - November 2009 EPFL/IEL/LTS4, Switzerland

• Research and Teaching assistant September 2004 - April 2009 EPFL/IEL/LTS4, Switzerland

• Research visitor January 2008 - April 2008 Redwood Center for Theoretical Neuroscience, University of California, Berkeley, US

• Student intern October 2001 - December 2001 EUPMT, Matar´o,Spain

RESEARCH PROJECTS • Developing a computational model for depth encoding in the visual cortex (ongoing work, jointly with B. A. Olshausen, UC Berkeley). The model will be tested on human data collected from the binocular eye tracker equipped with a stereo (experiments scheduled to begin in February 2011).

1 • Designed a method for sparse approximation of multi-view images (collaboration with P. Frossard, EPFL, and A. Ortega, USC)

• Designed a method for ultrasound tomography image reconstruction for breast cancer screening (collaboration with LCAV, EPFL and Karmanos Cancer Institute, Detroit, USA)

• Developed a method for overcomplete dictionaries for representing stereo images

• Developed an occlusion-resilient distributed coder for omnidirectional camera networks based on a novel geometric multi-view correlation model

• Designed and implemented a 3D objects coder based on the Spherical Matching Pursuit and the dictionary on the SO(3) group. Spherical shape representation exploited additionally for 3D face recognition

• Developed an interpolation method for spherical

• Developed a mammogram method for diagnosing breast cancer (at EUPMT, Matar´o,Spain)

TEACHING EXPERIENCE

• Teaching Assistant for the doctoral course Advanced processing, EPFL, fall 2004, fall 2005

• Teaching Assistant for the master course Image , EPFL, spring 2006

• Teaching Assistant for the bachelor course Digital Signal Processing (Traitement num´erique des signaux), teaching in french, EPFL, spring 2009

• Supervisor of master theses for three M.Sc. students at EPFL

• Gave lectures on Distributed coding within the Image Communications course, EPFL, spring 2007, 2008

GRANTS AND AWARDS

• Swiss National Science Foundation Fellowship 2010 Advanced researcher post-doctoral fellowship for 2011-2012

• International Conference on Image Processing 2010 Travel grant to attend the conference and present the work on Multi-view Matching Pursuit

• Swiss National Science Foundation Fellowship 2009 Prospective researcher post-doctoral fellowship for 2009-2010

• IBM T.J. Watson Research Center award 2008 ”Emerging Leaders in Multimedia 2008”, top 12 students world-wide

• EPFL Fellowship for the doctoral studies 2003 for the academic year 2003-2004

2 • The Certificate of Recognition of the University of Niˇs 2003 Faculty of Electronic Engineering Best Graduate Award

• Royal Family Karadjordjevic Award 2002 for the 100 best students in Serbia

• Norway Government Award 2001 for the 1000 best students in Serbia

PROFESSIONAL ACTIVITIES

• Organized a special session on Dictionary learning for sparse representation at ICASSP 2010 (jointly with P. Frossard, EPFL).

• Finance Chair for the 16th Packet Video Workshop, Lausanne, 2007.

• Technical Program Committee member for: ICME 2007, MMM 2009, 3DTV-CON 2009

• Reviewer for journals: IEEE Transactions on Image Processing, IEEE Transactions on Circuits and Systems for video technology, IEEE Transactions on Multimedia, IEEE Transactions on Medical Imaging, EURASIP Journal on Advances in Signal Processing, and Image Understanding, EURASIP Journal on Image and

• Reviewer for conferences: ICIP, ICASSP, EUSIPCO, NIPS, ICME, MMSP, ISCAS, Eurographics, MMM, 3DTV-CON, GlobeCom, PCS

INVITED TALKS

• Invited talks at international conferences

1. Wyner-Ziv coding of multi-view omnidirectional images with overcomplete decompositions, joint work with P. Frossard, invited at the IEEE International Conference on Image Processing, Atlanta (US), 2007 2. Coarse scene geometry estimation from sparse approximations of multi-view omnidirectional images, joint work with P. Frossard, invited at the European Signal Processing Conference, Poznan (Poland), 2007 3. A Geometrical Framework for Distributed Coding of Multiview Images, joint work with P. Frossard, invited at the DISCOVER workshop, Lisbon (Portugal), 2007 4. Balanced Distributed Coding of Omnidirectional Images, joint work with V.Thirumalai and P. Frossard, invited at the SPIE - Visual Communication and Image Processing Conference, San Jose (US), 2008 5. Distributed coding in camera networks with learned dictionaries, joint work with P. Frossard, invited to MobiMedia, International Mobile Multimedia Communications Conference, London (UK), 2009 6. Stereo dictionary learning for multiview scene representation, joint work with P. Frossard, invited to APSIPA Annual Summit and Conference, Sapporo (Japan), 2009

3 • Invited talks at institutes

1. Sparse approximations in signal processing and neuroscience, joint work with P. Frossard and B. A. Olshausen, invited at the Faculty of , University of Nis, Serbia, July 2010 2. Geometry-based representation of multi-view omnidirectional images, joint work with P. Frossard and A. Ortega, invited at Hewlett Packard Research Labs, Palo Alto (US), April 2010 3. Geometry-based distributed coding of multi-view images with sparse approximations, joint work with P. Frossard, invited at the Net/Comm/DSP Seminar Series, UC Berkeley (US), 2008 4. Coarse scene geometry estimation from multi-view omnidirectional images, joint work with P. Frossard, invited at the Heterogeneous Sensor Networks (HSN) Seminar Series, UC Berkeley (US), 2008 5. Geometry-based distributed coding of multi-view images with sparse approximations, joint work with P. Frossard, invited at the Emerging Leaders in Multimedia workshop, IBM Watson Research Center (US), 2008

LANGUAGES: English: fluent, French: fluent, Spanish: basic, Serbian: mother tongue

REFERENCES

• Prof. Pascal Frossard Signal Processing Laboratory LTS4, EPFL, Switzerland EPFL - FSTI - IEL - LTS4, Station 11, 1015 Lausanne, Switzerland P: +41 21 693 5655, F: +41 21 693 7600 Email: pascal.frossard@epfl.ch

• Prof. Antonio Ortega Signal and Image Processing Institute, Department of Electrical Engineering, University of Southern California 3740 McClintock Ave., EEB 436, Los Angeles, CA 90089-2564 P: (213) 740-2320, F: (213) 740-4651 Email: [email protected]

• Prof. Bruno A. Olshausen Redwood Center for Theoretical Neuroscience, Helen Wills Neuroscience Institute, UC Berkeley 575A Evans Hall, MC# 3198, Berkeley, CA 94720-3198 T: (510) 642-7250, F: (510) 642-7206 Email: [email protected]

• Prof. Pierre Vandergheynst Signal Processing Laboratory LTS2, EPFL, Switzerland EPFL - FSTI - IEL - LTS2, Station 11, 1015 Lausanne, Switzerland P: +41 21 693 5645, F: +41 21 693 7600 Email: pierre.vandergheynst@epfl.ch

4