Derivation of Image Characteristics to Describe the Appearance of Wood Flooring Products and Verification of Their Correspondence to Impressions Evaluated by Human Observers

Manufacturers of wood flooring products pursue enhancement of a wood-like feeling of their products. However, to numerically evaluate the effect of various treatments to improve the wood-like feeling has been very difficult. This study aims to develop a novel evaluation technique which objectively describes appearances of wood floorings which contribute to a wood-like feeling by certain image characteristics. Forty-seven flooring products were collected. Their surfaces were illuminated at various lighting azimuths, and high-resolution images of the surfaces were captured. The surface images were subjected to multi-resolution contrast analysis and correlation analysis. Four image characteristics of “micro visibility”, “macro visibility”, “contrast anisotropy”, and “gloss shift” were derived from these image analyses, which corresponded well with the reflectance properties of the flooring products. Additional 20 flooring products were prepared for sensory evaluation by 24 ordinary consumers. Multiple regression analysis revealed that the image characteristics computed for the 20 flooring products explained the results of sensory evaluation well.