Single-image Object Classification Based on Illuminette Construction from Shadow Imaging

Illuminettes, - bounded cross-sections of an illumination cone - are a property of an object with respect to an illumination source. This is intimately related to the use of cast shadows to classify and recognise the objects from their silhouette. This is since, from observation of unoccluded shadows on a terrain-like surface, illuminettes can be calculated. Here, we elaborate upon the criteria and assumptions necessary to perform illuminette-matching. We show how moments on these shadows may be used to perform matching making use of Hu moments. We also illustrate the impact on the Hu moments of Gaussian blurring on the illuminettes.