Evaluation criteria for appearance based maps

Authors:
Gorkem Erinc;Stefano Carpin
Affiliations:
University of California, Merced;University of California, Merced
Venue:
Proceedings of the 10th Performance Metrics for Intelligent Systems Workshop
Year:
2010

Citing 9
Cited 1

Robotic mapping: a survey

Exploring artificial intelligence in the new millennium
Epipole and fundamental matrix estimation using virtual parallax

ICCV '95 Proceedings of the Fifth International Conference on Computer Vision
Video Google: A Text Retrieval Approach to Object Matching in Videos

ICCV '03 Proceedings of the Ninth IEEE International Conference on Computer Vision - Volume 2
Distinctive Image Features from Scale-Invariant Keypoints

International Journal of Computer Vision
From images to rooms

Robotics and Autonomous Systems
Evaluating maps produced by urban search and rescue robots: lessons learned from RoboCup

Autonomous Robots
Image-based mapping and navigation with heterogenous robots

IROS'09 Proceedings of the 2009 IEEE/RSJ international conference on Intelligent robots and systems
Image-Based Visual Servoing for Nonholonomic Mobile Robots Using Epipolar Geometry

IEEE Transactions on Robotics
Fast and Incremental Method for Loop-Closure Detection Using Bags of Visual Words

IEEE Transactions on Robotics

Anytime merging of appearance-based maps

Autonomous Robots

Quantified Score

Hi-index	0.00

Visualization

Abstract

We present a set of task-based performance evaluation criteria designed to measure the quality of appearance based maps. Instead of aiming to measure a map's overall goodness, metrics defined in this paper focus on individual tasks, namely localization, planning, and navigation, and the quality of the map with respect to the their successful execution. The performance of a map in terms of localization is measured by the amount of information captured from the environment and the accuracy of this information. The planning metric favors instead maps with high connectivity and measures the validity of these connections. The navigation criterion, on the other hand, computes the robustness and stability associated with the paths that a robot will extract from the map. These metrics are tested on appearance maps created in our lab and their distinctiveness is shown.