art-synthetic

Turns a collection of hand-drawn notebook page photos into a structured dataset (transcribed text + style characteristics + artwork descriptions + spatial relationships), then uses that dataset to drive two generative outputs: a text-voice generator and an image-style generator.