Graininess characterization by multidimensional scaling

Abstract

A validation of the graininess attribute was made by means of a psychophysical experiment and the multidimensional scaling algorithm. A visual experiment was designed to obtain graininess differences to be used like the dissimilarity matrix in the MDS algorithm. The results revealed that two dimensions are needed to characterize the graininess effect. The BYK-mac-i instrument and a gonio-hyperspectral imaging system were employed to evaluate the statistical dimensions. On one hand, the first dimension correlated well with the graininess value provided by the BYK-mac-i (r2 = 0.9566). However, we were unable to find a relationship with dimension 2 and any parameter measured by this instrument. Furthermore, the images captured by the gonio-hyperspectral imaging system were processed. A good relationship with the correlation parameter was observed (r2 = 0.8958). However, no relationship was established with dimension 2. Based on these conclusions, further research is necessary that focuses on the new imaging processes and a new visual experiment.