Application of statistical theory to edge extraction in medical images
Conference Paper
Overview
Identity
Additional Document Info
View All
Overview
abstract
The authors describe an edge extraction algorithm specially developed for noisy images. It is based on the statistical theory of hypothesis testing and uses several parallel statistical tests in which indeterminate decisions are allowed to insure the reliability and reasonableness of the final decision. To demonstrate the capability of the new algorithm three images processed with this algorithm and with the Sobel edge operator for comparison are shown. The new algorithm is shown to work well on noisy data and requires no preprocessing. The new algorithm detects connected edge segments instead of individual edge points, which results in cleaner edges.