Skip to content

Svebor Karaman

  • PhD
  • Associate Research Scientist
  • Columbia University
  • 9h-indexImpact measure calculated using publication and citation counts. Updated daily.
  • 181CitationsNumber of citations received by Svebor's publications. Updated daily.

Recent publications

  • Learning discriminative and transformation covariant local feature detectors

    • Zhang X
    • Yu F
    • Karaman S
    • et al.
    Get full text
  • Strategies for multiple feature fusion with Hierarchical HMM: Application to activity recognition from wearable audiovisual sensors

    • Pinquier J
    • Karaman S
    • Letoupin L
    • et al.

Professional experience

Associate Research Scientist

Columbia University

April 2015 - Present

Postdoctoral research fellow

Universita degli Studi di Firenze

February 2012 - January 2015(3 years)

PhD student

Laboratoire Bordelais d'Informatique (LABRI)

October 2008 - December 2011(3 years)


PhD in Computer Science

Université Bordeaux 1

October 2008 - December 2011(3 years)

Master's Degree

Université Bordeaux 1

September 2007 - June 2008(9 months)

Engineer's Degree


September 2005 - June 2008(3 years)

Research interests

Machine learningPattern recognitionVideo indexingComputer vision


I am a French Computer Vision and Machine Learning researcher currently a Postdoctoral researcher in the DVMM Lab at Columbia University. Previously, I have spent three great years at the MICC (Media Integration and Communication Center) of the University of Florence in Italy. My research themes are image and video analysis. I am particularly interested in semantic concepts recognition in images and videos. I have made my PhD at the LaBRI – University of Bordeaux, under the supervision of Jenny Benois-Pineau and Rémi Mégret. During my PhD thesis, I have worked on human activities recognition by Hidden Markov Models (HMM) in videos recorded from a wearable device within the IMMED project. I have also developed an object recognition approach in the Bag-of-Visual-Words framework which integrates spatial information within semi-local features: the Graph-Words. I defended my PhD entitled “Indexing of Activities in Wearable Videos : Application to Epidemiological Studies of Aged Dementia” in 2011. Since joining the MICC in 2012, I have been highly involved in the MNEMOSYNE project. In this project multiple aspects of computer vision such as person detection, person tracking and re-identification are used to passively profile the interests of visitors in a museum to provide personalized multimedia content delivery. I am also still working on more general image and video classification problems.


Co-authors (37)

  • Lorenzo Seidenari
  • Giuseppe Serra
  • Marco Bertini
  • Aurélie Bugeau