A Computational Approach to Edge Detection
IEEE Transactions on Pattern Analysis and Machine Intelligence
Navigation and mapping in large-scale space
AI Magazine
DVI—a digital multimedia technology
Communications of the ACM
Communication in reactive multiagent robotic systems
Autonomous Robots
QuickTime VR: an image-based approach to virtual environment navigation
SIGGRAPH '95 Proceedings of the 22nd annual conference on Computer graphics and interactive techniques
Plenoptic modeling: an image-based rendering system
SIGGRAPH '95 Proceedings of the 22nd annual conference on Computer graphics and interactive techniques
Modeling visual attention via selective tuning
Artificial Intelligence - Special volume on computer vision
SIGGRAPH '96 Proceedings of the 23rd annual conference on Computer graphics and interactive techniques
Time-space tradeoffs for undirected graph traversal by graph automata
Information and Computation
A mobile robot that learns its place
Neural Computation
ECCV '96 Proceedings of the 4th European Conference on Computer Vision-Volume I - Volume I
Movie-maps: An application of the optical videodisc to computer graphics
SIGGRAPH '80 Proceedings of the 7th annual conference on Computer graphics and interactive techniques
Stochastic completion fields: a neural model of illusory contour shape and salience
ICCV '95 Proceedings of the Fifth International Conference on Computer Vision
Learning and Evaluating Visual Features for Pose Estimation
ICCV '99 Proceedings of the International Conference on Computer Vision-Volume 2 - Volume 2
Video Mosaics for Virtual Environments
IEEE Computer Graphics and Applications
| Hi-index | 0.00 |
For many tasks, we wish to record or recover thedescription of a remote environment so that it can be inspected by aperson. This is the problem we address in this paper. Rather thanrecovering a geometric description of an environment, as manyrobotics systems attempt to do, we seek to recover a model of anenvironment in terms of its appearance from a set of carefullyselected viewpoints. Our hope is that this type of model isboth more accessible to humans for many realistic tasks, and also morereadily achieved with automated systems. These viewpoints arelocations in the environment associated with views containing maximalvisual interest. This approach to environment representation isanalogous to image compression. Our goal is to obtain a set ofrepresentative views resembling those that would be selected by ahuman observer given the same task. Our computational procedure isinspired by models of human visual attention appearing in theliterature on human psychophysics. We make use of the underlyingedge structure of a scene, as it is largely unaffected by variationsin illumination. Our implementation uses a mobile robot to traversethe environment, and then builds an image-based virtualrepresentation of the environment, only keeping the views whoseresponses were highest. We demonstrate the effectiveness of ourattention operator on both single images, and in viewpoint selectionwithin an unknown environment.