Description
Photos preserve moments, but they also flatten them into fixed, two-dimensional frames. MemoAR proposes a different kind of memory archive, where moments take shape again, occupy space and return to everyday life. It is a gift for people who want their memories to exist beyond the frame.
MemoAR is a mobile AR app for iOS that turns existing photos and newly captured moments into interactive 3D memory objects. A childhood toy, a birthday cake or a favorite meal can be reconstructed from an image and brought into your surroundings through AR. Turning a photo into 3D changes more than its appearance: you can walk around a memory, see it from new angles, resize and reposition it, place it on real-world surfaces, interact with it using your hands, and bring several memories together in the same space. A cake from years ago can return to the dining table; something captured today can become a 3D memory you encounter again in the future. MemoAR also supports context-aware resurfacing, so meaningful memories reappear at relevant moments and places instead of staying buried in a photo library.
The team evaluated MemoAR with 16 participants across 329 memories. Participants created 3D memory objects not only to preserve meaningful moments, but also for aesthetics, collectibility, curiosity and tangible value. In AR, they used memories to reconnect past and present, make playful compositions, interact with remembered objects and reshape their surroundings. The result is a continuous memory loop: capture or rediscover a photo, turn it into 3D, place it and interact with it in AR, and encounter it again over time.
Before the event, MemoAR was a research concept with an early prototype, and the study design had IRB approval. During the event, the team rebuilt it into the current end-to-end iOS system: a polished frontend, a backend with a database and user management, the full photo-to-3D pipeline with Tripo, AR placement and manipulation including mid-air hand gestures, multi-object composition, location-based saving so memories can be left in a physical place and revisited there, and deployment of the whole system. The user study and its analysis were also done during the event.
How Tripo was used
Tripo is the core 3D generation engine behind MemoAR. Rather than using it for one-off asset creation, the team built the Tripo API into a full memory pipeline:
- Gemini identifies the meaningful subject and context in a photo and generates a stylized reference image.
- Tripo API image-to-3D turns the reference into a consistent 3D memory object.
- The models are then placed, manipulated, combined, saved in space and resurfaced over time in mobile AR.
