bild
School of
Computer Science
and Communication
Atsuto Maki
Research topics »
Publications »
Students »
Teaching »
Dataset »
About »
KTH / CSC / CVAP

Atsuto Maki, Professor

Address:Division of Robotics, Perception, and Learning (RPL)
Kungliga Tekniska Högskolan (KTH)
Teknikringen 14, plan 7
for mail use:SE - 100 44 Stockholm, SWEDEN
for ship use:SE - 114 28 Stockholm, SWEDEN

Research Interests: Machine Learning, Computer Vision, Interaction

Deep ConvNet Representations:

Motion Recognition & Detection:

  • Network Regularization - Feature Contraction
    • (demo code by Vladimir Li)
  • Factors of Transferability
  • A baseline for visual instance retrieval


    • Motion and Object Recognition:

    Motion Recognition & Detection:

  • Correlated Probabilistic Trajectories
  • A new distance for scale-invariant 3D shape recognition and registration
  • Demisting the Hough transform


    • Geotensity, Matching Measures:

    Geotensity, Matching measures: Geotensity, Matching measures:

  • Geotensity constraint: combining motion and lighting
  • SCVD: Sum of conditional variance of differences
  • Hyper-Patches for 3D Model Acquisition and Tracking
  • Learning-based stereo


    • Industrial Applications:

    Industrial Applications:

  • Subspace methods
  • Lighting environment estimation
  • Computational cameras


    • Clustering Techniques:

    Clustering techniques: Clustering techniques:

  • Random Forest Clustering
  • MCL-Superpixels: Compact Markov Cluster Algorithm for generating homogeneous superpixels


    • Attention, Biological & Machine Vision:

    Attention, Biological & Machine Vision::

  • Disparity Selection in Binocular Pursuit (Video)
  • Stereo Vision in Attentive Scene Analysis

    • Published by: Atsuto Maki (atsuto + at + kth.se)
    Updated 2020-10-15