INTRODUCING
Digital Twin Catalog
The world’s highest quality dataset for object reconstruction research.
WHAT IS IT?
A new dataset for 3D object reconstruction research
As more of our experience of the world moves into the digital realm, we increasingly rely on 3D digital twins of physical objects for e-commerce, entertainment, and advertisement. The majority of 3D objects today are manually created by technical artists at a high cost, creating a barrier to entry for many potential applications.
The Digital Twin Catalog (DTC) dataset is a large and highly detailed set of 3D object models that are sub-millimeter-level accurate with respect to their physical counterparts and highly photo-realistic. It also contains source capture data corresponding to different reconstruction techniques.
DTC will enable researchers to accelerate the development of new reconstruction techniques with the aim of making reconstruction ready for real-world applications.
DOWNLOAD THE DATASET BELOW
EXPLORE IN DATASET EXPLORER (Aria sequences only)
Dataset Content
- 2,400+ individually scanned 3D objects
- 400+ of the objects fully compatible with Aria Digital Twin Dataset
- Aria recordings of 100 of the objects
- DSLR camera captures of 50 of the objects
- Tools for downloading, loading, and visualization
Dataset Metrics
- 2,400+ 3D objects
- 200 Aria recordings
- 100 DSLR captures
WHAT IS IN THE DATASET?
2,400+ 3D Digital Twin Objects
2,400+ 3D objects captured using a state-of-the-art scanning rig and corresponding raw capture sequences.
Unlock new reconstruction research possibilities with the world’s most accurate 3D object models
We used a state of the art scanning rig to create 2,400+ geometrically detailed 3D object models with realistic PBR materials. The objects are diverse in category, geometry, and material and intended for researchers to improve their reconstruction algorithms.
Evaluate different types of reconstruction algorithms using Aria recordings and DSLR captures
For 100 of the objects that were scanned, we used Aria to create highly active recordings comprehensively capturing all angles of the objects, we also created passive walk-by recordings of those objects, and we used DSLR cameras to take surrounding images of 50 of the objects under two different lighting conditions. These each serve as evaluation data for egocentric and exocentric reconstruction techniques.
Object models fully compatible with Aria Digital Twin dataset
400+ of the scanned objects in the DTC dataset also correspond to scene objects from the Aria Digital Twin dataset and can be used along with Aria Digital Twin for object detection research.
DTC DATASET TOOLS
Comprehensive tools to load and visualize data easily
Accompanying tools to the DTC Dataset allow researchers to easily download the datasets, load the data, and visualize the data.
VIEW DTC DATASET TOOLS ON GITHUB
Object Reconstruction Academic Challenge
Academic Challenge Coming Soon!
Coming soon, we will be launching an academic competition for object reconstruction tasks using our dataset. In addition to the data components described above, we will be releasing 7,000+ highly detailed 3D objects with our partner, Shopify.
Access DTC Dataset and Accompanying Tools
If you are a researcher in AI or ML research, access the DTC Dataset and accompanying tools here.
By submitting your email and accessing the DTC Dataset, you agree to abide by the dataset license agreement and to receive emails in relation to the dataset.
Subscribe to Project Aria Updates
Stay in the loop with the latest news from Project Aria.
By providing your email, you agree to receive marketing related electronic communications from Meta, including news, events, updates, and promotional emails related to Project Aria. You may withdraw your consent and unsubscribe from these at any time, for example, by clicking the unsubscribe link included on our emails. For more information about how Meta handles your data please read our Data Policy.
