Digital image enhancement for in vivo laser scanning microscopy

Abstract

In vivo confocal laser scanning microscopy (CLSM) is a new method that provides skin images in horizontal plane at a level of resolution that allows to view microanatomic structures. This study examines whether certain digital image-processing steps can increase the visibility of various structures in CLSM. Fifty images were taken from normal skin of 25 probands, and 39 image enhancement procedures were created. Eight procedures that seemed to provide some quality enhancement were deliberately selected for further evaluation. Subsequently, a collection of random pairs of the original image and an image submitted to any of the eight selected procedures was rated by five independent observers. In three of the eight procedures tested, the modified image was significantly preferred to the original image (chi2-test,: P< or =0.001). In particular, smoothing, shading correction, delineate and grey-level normalization in various combinations were helpful in showing the characteristic honeycomb pattern, pigmented basal cell layer, cell borders and the nuclei more clearly. Digital image processing may help to increase visibility of in vivo CLSM images.