Concerto

Creates point-cloud embeddings learned jointly from 3D scans and 2D imagery.

Field
Cross-domain
Method
Foundation model
Data it takes
point-cloud, imagery
License
Apache-2.0 code, CC-BY-NC-4.0 weights — restricted use

Concerto trains on 3D scans and 2D images together, so its representations carry information from both. Where a survey has co-registered photography alongside lidar, that pairing is available to the model rather than discarded.

Like Sonata it is self-supervised, so it needs no labels to train, only labels for whatever task sits on top.

The code is Apache-2.0 but the weights are non-commercial.

Upstream project →

Catalog entry last checked 2026-08-19. All models →

Run Concerto on your data

The lab runs this on its own compute, in a container, with the inputs and parameters recorded alongside the result. Initial scoping conversations are free.

Contact the Lab