|dc.description.abstract||An algorithm, called the WSF algorithm, was developed which could predict the dyers‟ attributes of colour difference (depth, ΔD, brightness, ΔB, and hue, ΔH*) from the reflectance values of a pair of dyeings, enabling the dyer to take full advantage of colorimetric analysis. The algorithm was based on extensive experimental work to map surfaces of constant visual depth throughout the colour space and the thesis describes the methodology and the necessary calculations to determine the ΔD, ΔB and ΔH values of a pair of dyed samples. This algorithm was compared to other existing algorithms (the DBH and the Sato models) using two data sets with 49 dyed pairs for data set 1 and 117 dyed pairs for data set 2 respectively. The correlation of the values of ΔD, ΔB and ΔH determined using the WSF algorithm with the DBH and the Sato models showed an excellent relationship between these three algorithms for both the data sets. Qualitative comparison of the visual assessments of data set 2 with the WSF algorithm was encouraging but the quantitative comparison of the visual assessment for data set 1 was disappointing.
The pre-requisite of the WSF algorithm is the six equi-depth surfaces which have been defined numerically in the CIELAB colour space and previously reported as WSI depth surfaces. The first stage of this algorithm was to generate, using the WSI algorithm, the equi-depth line in the L* C* plane that passes through the L* C* coordinates of the standard. The K/S values of the batch were then iterated, until its depth became equal to that of the standard. At this point, the precise location of the batch on the equi-depth surface might be different from that of the standard. The linear distances between the batch and the standard, gave the differences in depth and brightness between the standard and the batch. A new approach was investigated for the hue correction of dyed pairs, where the hue of the batch was different from that of the standard. Real data of highest possible chroma values from Munsell colour atlas were used to create maximum chroma boundaries which were found necessary for the hue correction of the batch. It was noted that the DBH and the Sato models also incorporated hue correction of the batch.
The WSF algorithm described in Chapter 4 contained an iterative stage which created an additional complexity in programming. Therefore an alternative version of the WSF algorithm was developed, called the linear WSF model, which avoided the need for iteration and yielded the same results. This linear WSF algorithm strongly correlated with the WSF (iterative) model and also to other empirical models as well.||en_US