Color quantification in angioscopic video images.

Abstract

Colors in video representations of angioscopic images are up until now described by an human observer. Differences in settings of the monitor and the inherent poor ability of the human eye to classify colors objectively results in a very poor intraobserver as well as interobserver variability. A PC-based method is described to measure colors in a video image and to present the results in a novel C-diagram. Results with this method for standard calibrated colors are given. Possible sources of error are discussed and methods to minimize these errors are presented.