Department of Pattern Recognition

(eng): Two novel criteria capable of assessing spectral similarity and modelling plausibility of synthetic Bidirectional Texture Functions (BTF), static colours textures are presented. The criteria credibly compare their spectral contents. Their primary aim is to support optimal modelling algorithms and measurement setup development by comparing the originally measured target texture with its synthetic simulations. The suggested spectral similarity criteria are extensively tested on measured natural BTF textures and an artificial distinctly coloured texture, and favourably compared with several alternative spectral similarity criteria. The performance quality of the proposed criteria is demonstrated on a long series of specially designed monotonically spectrally degrading experiments.