A REPOSITORY OF INDOOR KNOWLEDGE
A REPOSITORY OF INDOOR KNOWLEDGE
We’re building a repository of indoor knowledge.
We’re building a repository of indoor knowledge.
We’re building a repository of indoor knowledge.
Structured, consent-native data from real indoor environments for physical AI. The spatial detail, context and variation that intelligent systems rarely have access to.
Structured, consent-native data from real indoor environments for physical AI. The spatial detail, context and variation that intelligent systems rarely have access to.
MISSION
MISSION
Physical intelligence needs a richer understanding of the places in which it will operate.
Physical intelligence needs a richer understanding of the places in which it will operate.
Physical intelligence needs a richer understanding of the places in which it will operate.
Most indoor environments remain absent from the data used to train machines. Our mission is to make that knowledge available, responsibly and at scale.
Most indoor environments remain absent from the data used to train machines. Our mission is to make that knowledge available, responsibly and at scale.
APPROACH
APPROACH
More data does not automatically create more understanding. We focus on six qualities that determine whether real-world data is genuinely useful.
More data does not automatically create more understanding. We focus on six qualities that determine whether real-world data is genuinely useful.
01
Fidelity
Preserve the physical detail that makes an environment valuable.
Preserve the physical detail that makes an environment valuable.
02
Information Gain
Seek what is missing, not simply more of what is already abundant.
Seek what is missing, not simply more of what is already abundant.
03
Relevance
Prioritise data with a clear relationship to the problem.
Prioritise data with a clear relationship to the problem.
04
Depth
Capture structure, relationships and context beyond appearance.
Capture structure, relationships and context beyond appearance.
05
Continuity
Understand environments as they change, not only as isolated snapshots.
Understand environments as they change, not only as isolated snapshots.
06
Reality
Begin with the world as it is: lived-in, variable and naturally complex.
Begin with the world as it is: lived-in, variable and naturally complex.
Each environment is grounded in consent and traceable provenance. Scale matters. What the data contains matters more.
Each environment is grounded in consent and traceable provenance. Scale matters. What the data contains matters more.
Every place contains knowledge.
Every place contains knowledge.
Every place contains knowledge.
We are building the infrastructure to capture it carefully, structure it faithfully and make it useful for physical intelligence.
We are building the infrastructure to capture it carefully, structure it faithfully and make it useful for physical intelligence.
TEAM
TEAM
SceneMe is being built by a small, focused team with backgrounds spanning large-scale augmented-reality systems and computer science at ETH Zurich.
SceneMe is being built by a small, focused team with backgrounds spanning large-scale augmented-reality systems and computer science at ETH Zurich.
Our work sits across technical systems and field operations, where dependable real-world data has to succeed both in software and in the field.
Our work sits across technical systems and field operations, where dependable real-world data has to succeed both in software and in the field.
We are building carefully, with a long view of what this repository can become.
We are building carefully, with a long view of what this repository can become.
Connect
Connect