Dr. Melih Kandemir

Heidelberg Collaboratory for Image Processing (HCI), Room B 108
Universität Heidelberg
Berlinerstr. 43 (Mathematikon Bauteil B)
D-69120 Heidelberg
Tel: +49-6221-54 14858
Email: name [dot] surname [at] iwr [dot] uni-heidelberg [dot] de

Background and Interests

I build probabilistic machine learning models for various applications including histopathology cancer diagnosis, diabetic retinopathy screening, and cell behaviour modeling from phase-contrast microscopy. My special focus is on Bayesian approaches to deep learning, multiple instance learning, structured learning, and active learning.

Google Scholar Profile: [WWW]



Gaussian processes for machine learning (Spring 2016) [WWW]
Bayesian modeling and inference (Spring 2014) [WWW]


  • Variational weakly supervised Gaussian processes
    M. Kandemir, M. Haußmann, F. Diego, K. Rajamani, J. van der Laak, F.A. Hamprecht
    BMVC, (2016), (Oral) [PDF] [Code]
  • Gaussian process density counting from weak supervision
    M. von Borstel, M. Kandemir, P. Schmidt, M. Rao, K. Rajamani, F.A. Hamprecht
    ECCV, (2016) [PDF]
  • The deep feed-forward Gaussian process: An effective generalization to covariance priors
    M. Kandemir, F.A. Hamprecht
    Journal of Machine Learning Research, (2015) [PDF] [Code]
    (SI: Feature Extraction: Modern Questions and Challenges)
  • Asymmetric transfer learning with deep Gaussian processes
    M. Kandemir
    ICML, Proceedings, (2015) [PDF] [Code][Talk]
  • Cell event detection in phase-contrast microscopy sequences from few annotations
    M. Kandemir, C.Wojek, F.A. Hamprecht
    MICCAI, Proceedings, (2015) [PDF]
  • Towards brain-activity-controlled information retrieval: Decoding image relevance from MEG signals
    J.P. Kauppi *, M. Kandemir *, V.M. Saarinen, L. Hirvenkari, L. Parkkonen, A.Klami, R. Hari, S. Kaski
    NeuroImage, (2015) * Equal Contributions [PDF]
  • Computer-aided diagnosis from weak supervision: A benchmarking study
    M. Kandemir, F.A. Hamprecht
    Computerized Medical Imaging and Graphics, (2015) [PDF] [Code]
  • Detection of retinopathy of prematurity using multiple instance learning
    P. Rani, E.R. Rajkumar, K. Rajamani, M. Kandemir, D. Singh
    ICACCI, (2015) [PDF]
  • Instance label prediction by Dirichlet process multiple instance learning
    M. Kandemir, F.A. Hamprecht
    UAI, Proceedings, (2014) [PDF] [Code]
  • Event detection by feature unpredictability in phase-contrast videos of cell cultures
    M. Kandemir, J.C. Rubio, U. Schmidt, C. Wojek, J. Welbl, B. Ommer, F.A. Hamprecht
    MICCAI, Proceedings, (2014) [PDF] [Code]
  • Empowering multiple instance histopathology cancer diagnosis by cell graphs
    M. Kandemir, C. Zhang, F.A. Hamprecht
    MICCAI, Proceedings, (2014) [PDF] [Code]
  • Multiple instance learning with response-optimized random forests
    C. Straehle, M. Kandemir, U. Koethe, F.A. Hamprecht
    ICPR, Proceedings, (2014) [PDF]
  • Multi-task and multi-view learning of user state
    M. Kandemir, A. Klami, M. Gönen, A. Vetek, S. Kaski,
    Neurocomputing, (2014) [PDF]
  • Digital Pathology: Multiple instance learning can detect Barrett's cancer
    M. Kandemir, A. Feuchtinger, A. Walch, F.A. Hamprecht,
    ISBI (Oral), Proceedings, (2014) [PDF]
  • Learning relevance from natural eye movements in pervasive interfaces
    M. Kandemir, S. Kaski,
    ICMI, Proceedings, (2012), [PDF]
  • Unsupervised inference of auditory attention from biosensors
    M. Kandemir, A. Klami, A. Vetek, S. Kaski,
    ECML, Proceedings, (2012), [PDF]
  • Multitask learning using regularized multiple kernel learning
    M. Gönen, M. Kandemir, S. Kaski,
    ICONIP, Proceedings, (2011), [PDF]
  • An Augmented reality interface to contextual information
    A. Ajanki, M. Billinghurst, H. Gamper, T. Järvenpää, M. Kandemir, S. Kaski, M. Koskela, M. Kurimo, J. Laaksonen, K. Puolamäki, T. Ruokolainen, T. Tossavainen
    Virtual Reality, (2011), [PDF]
  • Contextual information access with augmented reality
    A. Ajanki, M. Billinghurst, H. Gamper, T. Järvenpää, M. Kandemir, S. Kaski, M. Koskela, M. Kurimo, J. Laaksonen, K. Puolamäki, T. Ruokolainen, T. Tossavainen
    MLSP, Proceedings, (2010), [PDF]
  • Ubiquitous contextual information access with proactive retrieval and augmentation
    A. Ajanki, M. Billinghurst, M. Kandemir, S. Kaski, M. Koskela, M. Kurimo, J. Laaksonen, K. Puolamäki, T. Tossavainen
    IWUVR, Proceedings, (2010), [PDF]
  • Inferring object relevance from gaze in dynamic scenes
    M. Kandemir, V.-M. Saarinen, S. Kaski
    ETRA, Proceedings, (2010), [PDF]
  • Automatic segmentation of colon glands using object-graphs
    C. Gündüz-Demir, M. Kandemir, A.B. Tosun, C. Sokmensuer
    Medical Image Analysis, (2010), [PDF]
  • Object-oriented texture analysis for the unsupervised segmentation of biopsy images for cancer detection
    A.B. Tosun, M. Kandemir, C. Sokmensuer, C. Gündüz-Demir
    Pattern Recognition, (2008), [PDF]
  • A framework for real-time animation of liquid-rigid body interaction
    M. Kandemir, T. Capin, B. Ozguc
    CGI, Proceedings (2007), [PDF]


Learning mental states from biosignals
PhD Dissertation, (2013) , [PDF]

Segmentation of colon glands by object graphs
MSc Thesis, (2008) , [PDF]