The Narrative Interactive Immersion (NII) Framework is part of my ongoing PhD research, exploring how intangible heritage can be reinterpreted through game design principles. The project was presented at the Second Annual Review as a key milestone demonstrating the potential of immersive narrative systems for cultural heritage interpretation.

 

Stage 01 — 360° Site Capture and Point Cloud Generation

Footage is acquired on-site using Insta360 cameras mounted on an invisible stick, capturing the environment at three distinct heights — high level, eye level, and knee level — to ensure comprehensive spatial coverage. Raw video is then processed through FFmpeg at a 1:2 FPS sampling ratio to extract a controlled frame set. RealityCapture and Jawset Postshot are subsequently used to derive precise camera positions for each frame and generate the foundational 3D point cloud. Gaussian Splat training is then applied to this data, producing a photorealistic 3D representation of the site.

​

Stage 02 — Gaussian Splat Training 

The trained Gaussian Splat model serves as an interactive digital representation of a site, preserving both spatial geometry and material fidelity. Within the NII Framework, this representation functions as the immersive bridge between the physical heritage site and the simulation layer.

 

Stage 03 — Unreal Engine Third-Person Simulation

The digitised environment is imported into Unreal Engine and rendered as a navigable third-person simulation. Physical collision systems are integrated to enable spatially coherent movement through the heritage space, translating documentary capture into an embodied, game-engine-driven interpretive experience.

​