Use of the Laplacian of Gaussian operator in prostate ultrasound image processing

摘要

In computer analysis of prostate ultrasound images, detection of the contour of the prostate is difficult because of the ultrasound images' low resolution and high level of noise. The authors present a method for extracting the contour of the prostate employing the Laplacian of Gaussian (LoG) or Marr-Hildreth operator, the Gaussian kernel of which acts as a low pass filter eliminating the high frequency noise. LoG operators of various sizes have been tried on the ultrasound images and the results are encouraging.