Digital Reconstruction and Motion-Based Interactive Presentation of Qing Dynasty Court Opera Costumes

Qing Dynasty court opera costumes are material cultural heritage objects whose meanings are closely tied to performance, movement, and symbolic visual systems. While static digitization can document costume forms and surface details, it often fails to convey how motifs, sleeve structures, and bodily motion shape the performative experience of court opera. To address this limitation, we present an interactive digital system for the reconstruction and motion-based presentation of Qing Dynasty court opera costumes. The system combines: (1) high-fidelity 3D reconstruction of 30 representative costumes; (2) a FLUX-based diffusion generation module with LoRA fine-tuning for user-driven pattern exploration within the visual vocabulary of court opera costumes; and (3) a markerless pose-estimation pipeline using a single monocular camera to drive real-time costume animation and self-implemented lightweight hybrid costume dynamics in Unity. A comparative user study ( \(N=22\) ) evaluated the system against a static 3D display condition. Results indicate that the interactive system maintained comparable perceived costume-feature reproduction while significantly improving system usability and immersion ( \(p=0.0067\) ) and perceived cultural dissemination value ( \(p=0.0224\) ). These findings suggest that digital reconstruction, generative pattern exploration, and motion-based interaction can extend static costume archives into more engaging visual heritage experiences for objects whose cultural meanings depend on both material appearance and performative use.