Optimising the illumination spectrum for tissue texture visibility

Abstract

A light-emitting diode based spectrum optimisation is proposed to enhance the visibility of the texture of biological tissue. This optimisation method is based on maximising perceptual colour differences between pairs of colour patches using images of biological tissue. This approach has two advantages. First, by weighting the importance of colour differences, the impact of glint or specular reflection is reduced automatically. Second, this optimisation method puts the priority on small colour differences which could be more useful in enhancing the visibility of tissue texture. Furthermore, a psychophysical experiment has been conducted to evaluate the effectiveness of the proposed method. To be able to generalise our conclusions, different biological tissue types are used. The results show that illumination spectra that are optimised based on human perceptual colour differences significantly improve the visibility of tissue texture compared to illuminants such as CIE D65, and white LED light.

Improved and Robust Spectral Reflectance Estimation

“…The spectral reflectance of a surface is a unique property that is independent of the impinging illumination (Healey G 1991;Dana 2016), and is responsible, along with illumination's spectral power distribution (SPD), for the surface's (object's) color appearance. The spectral reflectance information is therefore useful for illuminating engineering applications (Durmus et al 2020), such as color tuning Durmus and Davis 2018), visual enhancement (Wang H et al 2018;Shen et al 2019), energy saving (Durmus and Davis 2015;Zhang JJ et al 2019), and computer vision application, such as surface/material recognition and characterization (Tominaga Shoji and Okajima 2000;Tu et al 2015), for image enhancement (Fu X et al 2015), for color constancy (Dixon and Shapiro 2017) and for geometry (shape) estimation from shading (Oxholm and Nishino 2016). Moreover, it is also useful in realistic material reproduction under a variety of illumination conditions in computer graphics (Filip et al 2017) and in relighting (Xing et al 2010), where multispectral reflectance approaches (Shrestha et al 2011;Khan et al 2013) can hardly meet the requirements.…”