14 November 2017 On correct evaluation techniques of brightness enhancement effect measurement data
Author Affiliations +
Abstract
This paper aims to establish confidence intervals of the quantification of brightness enhancement effects resulting from the use of pulsing bright light. It is found that the methods used so far may yield significant bias in the published results, overestimating or underestimating the enhancement effect. The authors propose to use a linear algebra method called the total least squares. Upon an example dataset, it is shown that this method does not yield biased results. The statistical significance of the results is also computed. It is concluded over an observation set that the currently used linear algebra methods present many patterns of noise sensitivity. Changing algorithm details leads to inconsistent results. It is thus recommended to use the method with the lowest noise sensitivity. Moreover, it is shown that this method also permits one to obtain an estimate of the confidence interval. This paper neither aims to publish results about a particular experiment nor to draw any particular conclusion about existence or nonexistence of the brightness enhancement effect.
© 2017 Society of Photo-Optical Instrumentation Engineers (SPIE)
Leoš Kukačka, Pascal Dupuis, Hideki Motomura, Jiří Rozkovec, Milan Kolář, Georges Zissis, Masafumi Jinno, "On correct evaluation techniques of brightness enhancement effect measurement data," Optical Engineering 56(11), 114103 (14 November 2017). https://doi.org/10.1117/1.OE.56.11.114103 . Submission: Received: 30 June 2017; Accepted: 23 October 2017
Received: 30 June 2017; Accepted: 23 October 2017; Published: 14 November 2017
JOURNAL ARTICLE
10 PAGES


SHARE
Back to Top