The Effects Of Lighting On The Visual Perception Of Surface Texture

Surface texture perception is highly influenced by the spectrum,direction,diffuseness and the complexity of the illumination.The perception of surface texture via light stimuli is importance for human life,because the surface texture of material has great impact on design and planning in relation to various products.However,texture perception is a complicated topic to investigate as it is also related to the interaction between material and lighting properties.We have investigated two classes of surface texture,namely 2D texture characterized by color non-uniformity or color variation of object surface which is largely dependent on the spectrum of the illumination,and 3D texture normally represented by the variation of surface height.Large amount research have been devoted on the relation between stimuli provided to the eye by reflected light,and the psychological perception of texture.However,there are still two questions unknown 1)for a certain illumination scene,what is best spectral power distribution which can provide maximum color non-uniformity;2)To what extent do the direction and diffuseness level of the illumination influence the perceived texture.These two questions are still not clear.So,in this thesis,we explored the above two questions from the following several aspects:1)We explored the relation between spectrum and 2D texture(the degree of color variation)based on a given illumination context,i.e.,surgical lighting.For this illumination context,an LED-based spectrum optimization was proposed,which aimed at enhancing texture visibility of biological tissue.This optimization method was based on maximizing perceptual color differences between pairs of color patches within the images of biological tissues in the visual field.In this case,texture visibility represents the degree of color non-uniformity or color variation.Firstly we measured the spectra reflectance of the biological tissues by multispectral imaging system,on which we used clustering method to find the relevant number of different color patches in the tissue.We then applied an optimization algorithm to make the mutual differences of these colors as large as possible giving different weights for the intensities of the 11 LEDs.2)Further to the first part of our work,we improved the optimization method by weighting the importance of perceptual color differences between pairs of color patches within the image,hence the impact of glint or specular reflection was reduced automatically.Besides,another metric min(1/?Eij)was proposed,which put the priority on small color differences which could be more useful to enhance texture visibility of tissues In addition,we evaluated different spectra optimization metrics by conducting a psychophysical experiment.To be able to generalize our conclusions,different biological tissue types are used.The results show that illumination spectra that are optimized based on human perceptual color differences significantly improve texture visibility compared to spectra that are optimized by other metric,or compared to standard illuminants such as CIE D65,and white LED light of 5500 K.3)Although the proposed metrics give better visibility compared to the other illuminations,they still may have some drawbacks.First,the proposed metrics based on maximizing color differences yield an illuminant with an SPD that is not optimized for color fidelity.It may provide surgeons relevant information on texture,but at the expense of color reliability.Hence,the optimized light probably cannot be used continuously,but should be combined with a light of high CRI.So,a method was explored to optimize the illumination spectrum for surgical lighting,which can simultaneously enhance the color contrast of biological tissue and the color rendering index(CRI)of the light source.Compared to conventional surgical lighting(i.e.based on halogen light sources),the optimized illumination has better color contrast and sufficiently high color rendering,i.e.both CIE Ra and R9 values are above 85.4)We explored how the direction and diffuseness of the illumination affect the perceived texture visibility.First,based on the light field theory and the definition of diffuseness,we generated different lighting conditions with 4 level directions and 4 level diffuseness.Second,two groups of materials were made with different roughness and specularity as we think the influence of lighting should be material-dependent.Third,the corresponding images of all prepared materials were captured under the generated lighting conditions.Then a psychophysical experiment was conducted to rate the amount of perceived textures illuminated by different light settings.