Evaluation of 50% acceptance level for shade sorting
Colourists are usually required to choose a colourdifference
value as a maximum permissible tolerance limit
for any given sample, against which to perform colour
acceptance determinations (shade passing). The same
condition should be considered in the shade sorting
process, particularly in the CCC process. Such a DE value
is used to determine whether the clustering process is
carried on or stopped during shade sorting.
Previous work on has been done on setting up equations
for colour acceptability through regression analysis, for
example by Vanderhoeven and Wilcock [18,19]. However,
no research has hitherto been aimed at systematically
studying a full range of colours to permit an allowable
tolerance limit for CCC shade sorting. In general, a 50%
rejection level is chosen to represent the maximum amount
of allowable failure. This represents the borderline case in
which colour could be regarded as a visual pass in one half
of the population and a visual fail in the another half. In
other words, it could be regarded as visual pass in a single
decision and visual fail in the next one.
In this study, the scattergrams obtained from the two
best-fit mathematical models, i.e. linear and logistical
models, and the two best-fit colour-difference equations, i.e.
CMC and CIE94, were selected for further analysis. The two
mathematical models imposed on the scattergram using the
two different equations were expressed in algebraic form,
as shown in Tables 3–6. Also the 50% rejection levels were
also calculated and these too are given.
Comparing the data in Tables 3–6 shows that the linear
and logistical analyses gave quite similar results with both
the CMC and CIE94 equations, so either could be chosen
for further application of the tolerance limit. One important
factor that needs to be mentioned is that the 50% rejection
level is only a recommended allowable failure rate. If