Inteligent Machine Perception

Intelligent Machine Perception

Our goal is to develop intelligent systems that can understand complex visual inputs, learn with minimal supervision and interact with dynamic, unstructured environments.

Research topics include

  • weakly supervised and unsupervised machine learning
  • visual recognition in images and videos 
  • embodied intelligence: perception, learning and motion generation for physical agents


Applications include video understanding and search, perception for robotics and autonomous driving, visual intelligence for augmented reality.

See for more details.

Contact Person

Josef Sivic