Edge Detection Algorithms in Nuclear Medicine

Abstract
Camera scans of phantoms and kidneys are computer processed to extract boundaries between areas of different densities. Several methods of edge detection are critically evaluated for their ability to determine boundaries in the presence of stochastic noise. A second order polynomial fit, utilizing the nearest neighbor method was found to be the most efficient method yielding the most consistent results for varying signal to noise ratios.

This publication has 3 references indexed in Scilit: