Options
1988
Book Article
Titel
What an autonomous robot can learn about its environment starting from scratch using image sequences
Abstract
We propose a method for the extraction of information from image sequences. No a priori knowledge is used except some very fundamental assumptions about the physical structure of the scene imaged onto the sensor. Experiments are demonstrated indicating the viability of the method. Moderately generalized models of visual objects are learned automatically and the direction of the moving sensor and the relative depth of the scene is extracted.
Language
English