The Design and Use of Steerable Filters
IEEE Transactions on Pattern Analysis and Machine Intelligence
An active vision architecture based on iconic representations
Artificial Intelligence - Special volume on computer vision
Real-time attention for robotic vision
Real-Time Imaging - Special issue on natural and artificial real-time imaging and vision
Optical normal flow estimation on log-polar images. A solution for real-time binocular vision
Real-Time Imaging - Special issue on natural and artificial real-time imaging and vision
Dynamic Vergence Using Log-Polar Images
International Journal of Computer Vision
A review of biologically motivated space-variant data reduction models for robotic vision
Computer Vision and Image Understanding
Efficient Encoding of Natural Time Varying Images Produces Oriented Space-Time Receptive Fields
Efficient Encoding of Natural Time Varying Images Produces Oriented Space-Time Receptive Fields
Detecting Perceptually Important Regions in an Image Based on Human Visual Attention Characteristic
Proceedings of the Joint IAPR International Workshop on Structural, Syntactic, and Statistical Pattern Recognition
A Humanoid Vision System for Versatile Interaction
BMVC '00 Proceedings of the First IEEE International Workshop on Biologically Motivated Computer Vision
A Feature-Driven Attention Module for an Active Vision System
Proceedings of the 24th DAGM Symposium on Pattern Recognition
A Method of Extracting Objects of Interest with Possible Broad Application in Computer Vision
BMCV '02 Proceedings of the Second International Workshop on Biologically Motivated Computer Vision
Active multi-view object search on a humanoid head
ICRA'09 Proceedings of the 2009 IEEE international conference on Robotics and Automation
Spatio-temporal attention mechanism for more complex analysis to track multiple objects
BVAI'05 Proceedings of the First international conference on Brain, Vision, and Artificial Intelligence
Hi-index | 0.00 |
The aim of this paper is to present our attempt in creating a visual system for a humanoid robot, which can intervene in nonspecific tasks in real-time. Due to the generic aspects of our goal, our models are based around human architecture. Such approaches have usually been contradictory, with the efficient implementation of real systems and its demanding computational cost. We show that by using PredN, a system for developing distributed real-time robotic applications, we are able to build a real-time scalable visual attention system. It is easy to change the structure of the system, or the hardware in order to investigate new models. In our presentation, we will also present our system with a number of human visual attributes, such as: log-polar retino-cortical mapping, banks of oriented filters providing a generic signature of any object in an image. Additionally, a visual attention mechanism - a psychophysical model - FeatureGate, is used in eliciting a fixation point. The system runs at frame rate, allowing interaction of same time scale as humans.