TU Darmstadt / ULB / TUbiblio

Saliency-based Gaze Direction Control for Active Vision

Mikhailova, Inna (2003):
Saliency-based Gaze Direction Control for Active Vision.
Technische Universitaet Darmstadt, Fachbereich Informatik, Fachgebiet Simulation und Systemoptimierung & Honda Research Institute Europe GmbH, [Diploma Thesis or Magisterarbeit]

Abstract

The research field of robotics exists now for over fifty years, but the robots are still not able to deal with simple tasks that can be easily accomplished by two year old children. Writing sophisticated algorithms, tuning the parameters, building extra-sensitive sensors results in demonstrations that are successful only in a confined experimental environment and fail in a natural one. How could robotics overcome this situation? In recent years researchers are more and more renouncing rigid algorithmic architectures that contain too much of explicit knowledge. These architectures are replaced with flexible ones, such as for example neural networks. It is surely the right way for coming closer to the performance of the human brain. Indeed, an autonomous robot needs structures that are adaptive, self-organised, able to learn and to interact with the environment. Following this philosophy we implement a gaze direction control for active vision.

Item Type: Diploma Thesis or Magisterarbeit
Erschienen: 2003
Creators: Mikhailova, Inna
Title: Saliency-based Gaze Direction Control for Active Vision
Language: English
Abstract:

The research field of robotics exists now for over fifty years, but the robots are still not able to deal with simple tasks that can be easily accomplished by two year old children. Writing sophisticated algorithms, tuning the parameters, building extra-sensitive sensors results in demonstrations that are successful only in a confined experimental environment and fail in a natural one. How could robotics overcome this situation? In recent years researchers are more and more renouncing rigid algorithmic architectures that contain too much of explicit knowledge. These architectures are replaced with flexible ones, such as for example neural networks. It is surely the right way for coming closer to the performance of the human brain. Indeed, an autonomous robot needs structures that are adaptive, self-organised, able to learn and to interact with the environment. Following this philosophy we implement a gaze direction control for active vision.

Place of Publication: Technische Universitaet Darmstadt
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Simulation, Systems Optimization and Robotics Group
Date Deposited: 26 Jun 2019 11:50
Export:
Suche nach Titel in: TUfind oder in Google

Optionen (nur für Redakteure)

View Item View Item