Universitšt Bonn: Autonomous Intelligent Systems Group   Computer Science Institute VI: Autonomous Intelligent Systems

Deep Learning Group

The Deep Learning Group is focused on Computer Vision with emphasis on object categorization, object-class segmentation, and action recognition tasks. Here Deep Learning means learning using hierarchical methods such as neural networks, or deep graphical models. To make calculations on full sized images feasible, we make use of the Nvidia CUDA framework, to perform calculations on consumer graphic cards.

Our research on Learning Hierarchial Invariant Features for Action Recognition is supported by the DFG Research Training Group 1572 Bionics.

Horizontal and Vertical Dependencies in Images    Object class segmentation: Inria Graz-02 data set  Semantic segmentation of RGB-D images

  LabelMe-12-50k Object Recognition Data Set

Convolutional Neural Network                     Local-Impact RBM (LIRBM)

Universitšt Bonn, Institute for Computer Science, Departments: I, II, III, IV, V, VI