Share Email Print
cover

Optical Engineering

Parametric and nonparametric edge detection for speckle degraded images
Author(s): Kevin D. Donohue; Mohammad Rahmati; Laurence G. Hassebrook; P. Gopalakrishnan
Format Member Price Non-Member Price
PDF $20.00 $25.00

Paper Abstract

Edge detection is useful for locating objects of interest within an image and reducing the amount of processing required for the image analysis. Methods for edge detection usually involve convolving the image with an operator designed to have a relatively high output when an edge or gradient is present. In textured or noisy images, however, the edge-detection scheme must take into account the nature of random fluctuations throughout the image to limit erroneous detections. Two statistical tests for detecting edges in images corrupted by speckle are presented. The tests are based on the nonparametric Wilcoxon two-sample test and a parametric test derived from an exponential model for the speckle. These edge detectors are presented as null hypothesis tests for determining the presence of an edge based on significant changes in the location parameters (first-order statistics) between pixel neighborhoods. The null hypothesis test formulation allows for threshold determination based on desired false-alarm probabilities. Simulation results demonstrate the ability of the nonparametric test to maintain a constant false-alarm probability under variations in the skewness of the speckle statistics, whereas superior detection probabilities are achieved with parametric tests over a broad range of statistical variations. Examples of detector performance for ultrasonic images from breast tissue are also presented and interpreted in terms of the simulation results. Conclusions are presented outlining conditions for the successful application of parametric and nonparametric techniques for edge detection using first-order statistics.

Paper Details

Date Published: 1 August 1993
PDF: 12 pages
Opt. Eng. 32(8) doi: 10.1117/12.143717
Published in: Optical Engineering Volume 32, Issue 8
Show Author Affiliations
Kevin D. Donohue, Univ. of Kentucky (United States)
Mohammad Rahmati, Univ. of Kentucky (United States)
Laurence G. Hassebrook, Univ. of Kentucky (United States)
P. Gopalakrishnan, Univ. of Kentucky (United States)


© SPIE. Terms of Use
Back to Top