Comparative Study of Metrics for Spectral Match Quality

The selection of metrics for spectral matches is fundamental to MVSI (multi-channel visible-spectrum imaging) otherwise known as spectral imaging. The metrics used for spectral matches can impact everything from the selection of the filters used for multi-channel capture to the evaluation of the spectral estimation. However, there is, as yet, no consensus on which metric should be applied for spectral matches. The purpose of this research is to compare various metrics that have been used for spectral matches. The metrics for spectral comparison were categorized in four classes: CIE color difference equations, spectral curves difference metrics, metamerism indices and weighted spectral metrics. Here we show an analysis of the appropriateness and weakness of each metric. We compare their use for various types of spectral mismatches resulting from problems in imaging calibration, out-of-gamut colors and those due to metamerism.