Generative study of form, redundancy, and subject identity
SHUM investigates decorative excess as a method for transforming familiar objects. Beads, pendants, fittings, and various textures cease to be mere additions to an object, becoming its primary defining characteristic.
The central research question is: what remains of an object when its original form begins to disappear, but the decorative elements persist?
The project hypothesis is that a generative model, fine-tuned on images of a single signature object, will be able to extract its stable visual characteristics and reproduce them in new variations. It is assumed that the neural network will assimilate not only the general silhouette of the vest but also the principles governing the arrangement of decorative elements—the heterogeneity of the beads, the combination of materials, the suspension system, color contrasts, and characteristic visual density.
