Chromatic distortion during angioscopy: assessment and correction by quantitative colorimetric angioscopic analysis.
Redirect to Publisher's version
(Publisher's version.url.txt, 79 bytes)
Angioscopy represents a diagnostic tool with the unique ability of assessing the true color of intravascular structures. Current angioscopic interpretation is entirely subjective, however, and the visual interpretation of color has been shown to be marginal at best. The quantitative colorimetric angioscopic analysis system permits the full characterization of angioscopic color using two parameters (C1 and C2), derived from a custom color coordinate system, that are independent of illuminating light intensity. Measurement variability was found to be low (coefficient of variation = 0.06-0.64%), and relatively stable colorimetric values were obtained even at the extremes of illumination power. Variability between different angioscopic catheters was good (maximum difference for C1, 0.022; for C2, 0.015). Catheter flexion did not significantly distort color transmission. Although the fiber optic illumination bundle was found to impart a slight yellow tint to objects in view (deltaC1 = 0.020, deltaC2 = 0.024, P < 0.0001) and the imaging bundle in isolation imparted a slight red tint (deltaC1 = 0.043, deltaC2 = -0.027, P < 0.0001), both of these artifacts could be corrected by proper white balancing. Finally, evaluation of regional chromatic characteristics revealed a radially symmetric and progressive blue shift in measured color when moving from the periphery to the center of an angioscopic image. An algorithm was developed that could automatically correct 93.0-94.3% of this error and provide accurate colorimetric measurements independent of spatial location within the angioscopic field. In summary, quantitative colorimetric angioscopic analysis provides objective and highly reproducible measurements of angioscopic color. This technique can correct for important chromatic distortions present in modern angioscopic systems. It can also help overcome current limitations in angioscopy research and clinical use imposed by the reliance on visual perception of color.
- Image Processing, Computer-Assisted
- *Image Interpretation, Computer-Assisted