Kawasemi-Colours: an open end-to-end platform for structure-colour design based on avian structural colour

Abstract

Structural colour provides durable, pigment-free colouration by controlling the interaction of light with nanostructured materials rather than by using wavelength-selective pigments. In birds, non-iridescent blue and green plumage arises from coherent scattering by quasi-ordered keratin-air nanostructures, making these natural systems attractive models for sustainable colour engineering. Despite decades of research on their optical mechanisms, open end-to-end implementations that connect biological nanostructures with colour prediction, inverse design and physics-based evaluation remain limited. Here I present Kawasemi Colours, an open end-to-end platform for structure-colour design based on avian structural colour. From published electron micrographs, I assembled a provenance-tracked dataset of spongy, channel and quasi-ordered sphere nanostructures with associated colour metadata. I extracted interpretable structural descriptors, developed compact forward models that predict colour from nanostructure images, and implemented an analytic inverse design tool that generates candidate structures for a target colour. Both directions are provided through self-contained interactive web applications. Kawasemi Colours provides an open framework for studying and designing structural colour from biological nanostructures. By integrating curated data, interpretable machine learning, wave optics, and interactive design tools, the platform lowers the barrier to comparative structural-colour research and provides a cumulative foundation for testing and refining structure-colour relationships as new biological data become available.