Effects of Image Processing Using Honeycomb-Removal and Image-Sharpening Algorithms on Visibility of 27-Gauge Endoscopic Vitrectomy

Abstract

Endoscopic vitrectomy with small gauge probes has clinical potentials, but intraocular visibility is inherently limited by low resolution and dim illumination due to the reduced number of optic fibers. We investigated whether honeycomb-removal and image-sharpening algorithms, which enable real-time processing of live images with a delay of 0.004 s, can improve the visibility of 27-gauge endoscopic vitrectomy. A total of 33 images during endoscopic vitrectomy were prepared, consisting of 11 original images, 11 images after the honeycomb-removal process, and 11 images after both honeycomb-removal and image-sharpening procedures. They were randomly presented to 18 vitreous surgeons, who rated each image on a 10-point scale. The honeycomb-removal algorithm almost completely suppressed honeycomb artifacts without degrading the background image quality. The implementation of image-sharpening algorithms further improved endoscopic visibility by optimizing contrast and augmenting image clarity. The visibility score was significantly improved from 4.27 ± 1.78 for the original images to 4.72 ± 2.00 for the images after the honeycomb-removal process (p < 0.001, linear mixed effects model), and to 5.40 ± 2.10 for the images after both the honeycomb-removal and image-sharpening procedures (p < 0.001). When the visibility scores were analyzed separately for 10 surgeons who were familiar with endoscopic vitrectomy and 8 surgeons who were not, similar results were obtained. Image processing with honeycomb-removal and image-sharpening algorithms significantly improved the visibility of 27-gauge endoscopic vitrectomy.