A hierarchical filter scheme for efficient corner detection

Authors:
Tobias Stammberger;Markus Michaelis;Maximilian Reiser;Karl-Hans Englmeier
Affiliations:
Institut for Medical Informatics, GSF National Research Center for Environment and Health, Ingolstädter Landstr. 1, D-85764 Oberschleiβheim, Germany and Institut für Radiologische D ...;Institut for Medical Informatics, GSF National Research Center for Environment and Health, Ingolstädter Landstr. 1, D-85764 Oberschleiβheim, Germany;Institut für Radiologische Diagnostik, Klinikum Groβhadern, Marchioninistr. 15, D-81337 München, Germany;Institut for Medical Informatics, GSF National Research Center for Environment and Health, Ingolstädter Landstr. 1, D-85764 Oberschleiβheim, Germany
Venue:
Pattern Recognition Letters
Year:
1998

Citing 9
Cited 1

Corner characterization by differential geometry techniques

Pattern Recognition Letters
The Geometry of Differential Operators with Application to Image Processing

IEEE Transactions on Pattern Analysis and Machine Intelligence
Grey level corner detection: a generalization and a robust real time implementation

Computer Vision, Graphics, and Image Processing
Corner detection

Pattern Recognition
Generic Neighborhood Operators

IEEE Transactions on Pattern Analysis and Machine Intelligence
A computational approach for corner and vertex detection

International Journal of Computer Vision
Using partial derivatives of 3D images to extract typical surface features

Computer Vision and Image Understanding
A lie group approach to steerable filters

Pattern Recognition Letters
Gray-level corner detection

Pattern Recognition Letters

Behavior of the Laplacian of Gaussian Extrema

Journal of Mathematical Imaging and Vision

Quantified Score

Hi-index	0.10

Visualization

Abstract

There are a number of differential geometric based corner detectors in the literature that calculate the differential characteristics of the image intensity surface. These operators require 5 or 9 convolutions with derivative kernels in 2D or 3D respectively, what is expensive in terms of time and memory requirements. In this paper we propose an efficient approach to calculate the response of these operators. We introduce a set of orthogonal second-order Gaussian derivative kernels. The whole image is convolved with only one of these kernels. The application of the remaining kernels is restricted to the neighborhoods of positions where this kernel shows high local activity in the image.