Extracting Slow Subspaces from Natural Videos Leads to Complex Cells

Natural videos obtained from a camera mounted on a cat’s head are used as stimuli for a network of subspace energy detectors. The network is trained by gradient ascent on an objective function defined by the squared temporal derivatives of the cells’ outputs. The resulting receptive fields are invariant to both contrast polarity and translation and thus resemble complex type receptive fields.

• Type: Conference
• Year: 2001
• Where: ICANN
• Authors: Christoph Kayser, Wolfgang Einhäuser, Olaf Dümmer, Peter König, Konrad P. Körding
• Keywords: ICANN 2001, Neural Networks, Receptive fields, Squared Temporal Derivatives, Subspace Energy Detectors