Error propagation analysis in color measurement and imaging

✍ Scribed by Peter D. Burns; Roy S. Berns

• Publisher: John Wiley and Sons
• Year: 1997
• Tongue: English
• Weight: 202 KB
• Volume: 22
• Category: Article
• ISSN: 0361-2317
• DOI: 10.1002/(sici)1520-6378(199708)22:4<280::aid-col9>3.0.co;2-l

✦ Synopsis

We apply multivariate error-propagation analy-Uncertainty or noise in a detected or recorded color signal can arise from many sources, e.g., detector dark sis to color-signal transformations. Results are given that indicate how linear, matrix, and nonlinear transforma-current, exposure shot noise, calibration variation, or varying operating conditions. If a physical model of the system tions influence the mean, variance, and covariance of color-measurements and color-images. Since many signal and its associated signal processing is available, the influence of various sources on system performance can be processing paths include these steps, the analysis is applicable to color-measurement and imaging systems. Ex-understood for both color-measurement 1-8 and imaging applications. [11][12] In addition, statistical aspects of human vi-pressions are given that allow image noise or error propagation for a spectrophotometer, colorimeter, or digital sion can also be explicitly included in the analysis. 13 This approach allows the comparison of design/technology camera. In a computed example, error statistics are propagated from tristimulus values to CIELAB coordinates.

choices in terms of system performance requirements, e.g., color error or signal-to-noise ratio. We consider the case The resulting signal covariance is interpreted in terms of CIELAB error ellipsoids and the mean value of color-of general stochastic error sources, which can be functions of exposure level, wavelength, etc. difference measures, DE* ab and DE* 94 . The application of this analysis to system design is also illustrated by relat-Measurements of systematic error are often used to ing a DE* 94 tolerance to equivalent tristimulus-value error evaluate accuracy during system calibration. Methods of statistics.