Learning Saccadic Gaze Control via Motion Prediction

Per-Erik Forssén
CRV07, Montréal, Québec, Canada
4th Canadian Conference on Computer and Robot Vision
Pages 44-51
May 2007

Abstract

This paper describes a system that autonomously learns to perform saccadic gaze control on a stereo pan-tilt unit. Instead of learning a direct map from image positions to a centering action, the system first learns a forward model that predicts how image features move in the visual field as the gaze is shifted. Gaze control can then be performed by searching for the action that best centers a feature in both the left and the right image. By attacking the problem in a different way we are able to collect many training examples in each action, and thus learning converges much faster. The learning is performed using image features obtained from the Scale Invariant Feature Transform (SIFT) detected and matched before and after a saccade, and thus requires no special environment during the training stage. We demonstrate that our system stabilises already after 300 saccades, which is more than 100 times fewer than the best current approaches.

Full Paper

Portable document format file PDF
On-line proceedings available on the IEEE Explore website.

Bibtex entry

@InProceedings{f07,
  author = 	 {Per-Erik Forss{\'e}n},
  title = 	 {Learning Saccadic Gaze Control via Motion Prediction},
  booktitle = 	 {4th Canadian Conference on Computer and Robot Vision},
  pages = 	 {44-51},
  year = 	 {2007},
  address = 	 {Montr{\'e}al, Qu{\'e}bec, Canada},
  month = 	 {May},
  publisher =    {{IEEE} Computer Society}
}

Other publications from the LCI lab can be found here.