Title :
Detection and labeling of retinal vessels for longitudinal studies
Author :
Wood, Sally L. ; Qu, Gongyuan ; Roloff, L.W.
Author_Institution :
Santa Clara Univ., CA, USA
Abstract :
Images of the retina taken as a routine opthalmologic procedure can provide early indications of damage to the retinal nerve fiber layer. Automatic processing of such images for screening is hindered by variable image acquisition and film processing parameters, and interference from the vessel structure in the image. This papers presents procedures for equalizing the image variability so that nonlinear morphological filtering methods can be used to locate and model vessel segments. Frequency domain analysis can then be used to assess texture parameters in the areas which do not include vessels
Keywords :
biomedical imaging; eye; feature extraction; filtering theory; frequency-domain analysis; image texture; mathematical morphology; medical image processing; neurophysiology; photographic applications; vision defects; automatic image processing; disease; early damage indications; film processing parameters; frequency domain analysis; glaucoma; image acquisition; image variability; longitudinal studies; nonlinear morphological filtering methods; photographic images; retina images; retinal nerve fiber layer; retinal vessel detection; retinal vessel labeling; routine opthalmologic procedure; screening; texture parameters; vessel segments; Biomedical optical imaging; Diseases; Image quality; Interference; Labeling; Lenses; Nerve fibers; Optical films; Retina; Retinal vessels;
Conference_Titel :
Image Processing, 1995. Proceedings., International Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-8186-7310-9
DOI :
10.1109/ICIP.1995.537606
