Algorithmic Flesh
Interactive installation
"Algorithmic Flesh" is an interactive multimedia installation that reimagines the UNESCO Dunhuang Apsara murals through a novel "Somatic Prompting" paradigm. Moving beyond static digital preservation, this solo project translates empirical dance research into a real-time, zero-latency computer vision pipeline. Driven by customized generative AI workflows and YOLO 26, the system probabilistically maps participants' embodied kinetics—such as the precise "Acolyte Worshipping Moon" pose—against an archive of ancient postures to hallucinate a "post-human choreography." This co-interpretive framework bridges biological flesh and digital algorithms, transforming cultural heritage into a visceral, evolving physical interaction.
Algorithmic Flesh: An Embodied AI Co-Interpretive Ritual for Revitalizing Dunhuang Apsara Murals
"Algorithmic Flesh" tackles the 'Crisis of Stasis' in the digital preservation of cultural heritage. Focusing on the Dunhuang Apsara murals, this interactive installation proposes an alternative interaction paradigm termed "Somatic Prompting." Instead of conventional inputs, users interact through real-time embodied kinetics, breathing life into historical archives.
The project’s genesis is rooted in a sustained, empirical methodology. The foundational Dunhuang dance vocabulary was painstakingly reconstructed through decades of somatic practice and biomechanical analysis of the murals. Translating this organic knowledge into a machine-readable format required rigorous iterative refinement. We developed a zero-latency computer vision pipeline driven by a customized YOLO 26 model and MediaPipe. Furthermore, our customized generative AI workflows orchestrate a precise topological mapping between the contemporary body and archival poses. For instance, detecting specific kinetic anchors like the "Acolyte Worshipping Moon" pose triggers a cascade of targeted generative responses, ensuring fluid movement without breaking the algorithmic logic.
Crucially, the generated movements are not mechanical restorations. Through "Speculative Fabulation," the system hallucinates inferential reproductions—a "post-human choreography" shaped by the methodological friction between contemporary biological flesh and ancient aesthetic laws. When a participant achieves a state of "Dharmic Sync"—evaluated via a "Nine-Grade" logic recognizing specific poses and hand gestures (mudras)—the system dynamically generates philosophical texts from ancient Dunhuang manuscripts.
As a solo endeavor presenting a non-Western cultural computing framework, "Algorithmic Flesh" serves as an experimental bridge spanning computer science, digital art, and history. It demonstrates how rigorous technological experimentation and somatic engagement can transform static artifacts into living, co-interpretive performances, proving that cultural heritage is an evolving, speculative algorithm of the present.