Concept Statement

We remember in places. A meal belongs to the desk where it was eaten. Rain belongs to the window where it was watched. Memory is not a collection of images filed away in the mind; it is spatial, anchored to the rooms and surfaces where life happened. We know this intuitively, but we have never had a medium that lets someone else feel it.

3D Gaussian Splatting (3DGS) can preserve not just images of a space, but the space itself, as a cloud of millions of particles that hold the color, light, and depth of a real place. Combined with VR, this means a person can physically walk through a space that no longer exists, and discover the memories that lived there.

My thesis project is a VR experience built inside a 3DGS scan of the apartment I lived in for a year in Brooklyn, New York. It was my first home outside of Taiwan, and it no longer exists as I knew it. Over that year, I captured hundreds of photos and videos from inside the apartment: meals at my desk, light through the window across seasons, the room transforming from an empty shell into a lived-in space.

In the VR experience, the viewer enters the apartment and discovers memory fragments (photos, videos) emerging from the locations where they happened. The desk held hundreds of meals over the year, so food photos surface one by one when the viewer stands there and looks down. The window framed the street across seasons, so the view appears when the viewer looks out. The interaction is gaze-triggered: walk close and look the way you naturally would, and the memory surfaces. Turn away, and it fades. You discover memories by being present in the same spot where they were made.

The 3D space is a container for 2D memories. Photos and videos exist inside the navigable room the same way a photo album sits on a shelf: the space holds them, gives them context, and places them where they belong. A new dimension does not replace the previous one. A 3D scanned space does not replace the flat images captured inside it. It holds them.

When a viewer focuses on a memory, the surrounding space softens. The 3DGS particles spread apart, edges blur, the room becomes impressionistic. When the viewer looks away from the memory and back at the space, it resolves into clarity again. The space breathes between two states: high-fidelity and particle abstraction. This mirrors the phenomenology of remembering: when you are deep in a specific memory, your awareness of the present space fades; when you return to the present, the world sharpens.

This project draws on my background as a theatrical lighting designer. For years, I have used space, light, and atmosphere to transmit feeling to an audience. Theater is ephemeral: the light disappears when the show ends. 3DGS and VR offer a way to preserve and share spatial experience beyond the live moment. This is not a study of the technology. It is the continuation of the same practice: designing how someone moves through a space, what they encounter, and what they feel.

The core question is experiential: can a viewer feel the spatial nature of memory by physically navigating a captured personal space in VR? And can 3DGS and VR serve as a medium for preserving and sharing spatial memory?

Keywords

experience, memory, space, interaction, preservation

Thesis Statement

My thesis explores whether a VR experience built inside a 3DGS-captured personal space can let viewers feel the spatial nature of memory, by discovering location-triggered memory fragments through gaze and physical navigation, and experiencing the space shift between fidelity and abstraction in response to their attention.