An Efficient Data Structure for Feature Extraction in a Foveated Environment

Authors:
Efri Nattel;Yehezkel Yeshurun
Affiliations:
-;-
Venue:
BMVC '00 Proceedings of the First IEEE International Workshop on Biologically Motivated Computer Vision
Year:
2000

Citing 6
Cited 0

Dynamic aspects in active vision

CVGIP: Image Understanding - Special issue on purposive, qualitative, active vision
Space variant image processing

International Journal of Computer Vision
Machine vision

Machine vision
Context-free attentional operators: the generalized symmetry transform

International Journal of Computer Vision - Special issue on qualitative vision
An active foveated vision system: attentional mechanisms and scan path convergence measures

Computer Vision and Image Understanding
On the Advantages of Polar and Log-Polar Mapping for Direct Estimation of Time-To-Impact from Optical Flow

IEEE Transactions on Pattern Analysis and Machine Intelligence

Quantified Score

Hi-index	0.01

Visualization

Abstract

Foveated sampling and representation of images is a powerful tool for various vision applications. However,there are many inherent difficulties in implementing it. We present a simple and efficient mechanism to manipulate image analysis operators directly on the foveated image; A single typed table-based structure is used to represent various known operators. Using the Complex Log as our foveation method, we show how several operators such as edge detection and Hough transform could be efficiently computed almost at frame rate, and discuss the complexity of our approach.