The Importance of Smoothness Constraints on Spectral Object Reflectances When Modeling Metamer Mismatching

Tarek Stiebel, Dorit Merhof; Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2017, pp. 2985-2992

This paper analyzes the influence of multi-spectral imag- ing onto the severity of metamer mismatching, in partic- ular in the context of color-accuracy of machine vision. Camera signals associated with simulated as well as real world multi-spectral imaging systems when viewing differ- ent objects under different lighting conditions were calcu- lated. Based on the calculated camera signals, the associ- ated MMBs were computed when changing towards the CIE standard observer under illuminant D65. The results show that an increased number of channels used in multi-spectral imaging systems do not necessarily de- crease the severity of metamer mismatching. However, it is also shown that this is due to the limited capabilities of current image acquisition models which are not able to cor- rectly compute a realistic MMB as they neglect any smooth- ness constraints on spectral object reflectances.