Unfinished Decay: Animating Historical Images from Gigapixel Digitizations via Fragment-Based AI Workflows 40

Unfinished Decay is an AI-assisted animation project that explores how historical gigapixel images of cultural heritage artworks can be reinterpreted through contemporary generative image-to-video and video-to-video systems. Building on prior research from Brosch AI – Distorted Dreams (2025) this work shifts focus from stylistic distortion toward material-based interpretation. Using a gigapixel scan of Red Dinosaurs (1926), an unfinished late painting by Austrian artist Klemens Brosch, the project animates microscopic surface features – dust, cracks, pigment erosion – as primary carriers of meaning. Current AI animation tools operate under strict resolution constraints (typically 1–4K), limiting their applicability to ultra-high-resolution cultural artifacts. To address this limitation, Unfinished Decay (2026) introduces a fragment-based workflow that extracts micro-regions from gigapixel imagery, applies AI-driven motion, and recomposes the results into a cohesive cinematic structure. The resulting five-minute film unfolds as a silent visual essay on decay, fragility, and temporal persistence, positioning AI as a speculative collaborator in the artistic reinterpretation of cultural heritage.