Skip to main content
smithsonian collections sketch

Smithsonian 3D Metadata Model

By DPO on Thu, 11/01/2018 - 11:33

Introduction
As 3D digitization becomes more common in collections documentation, there is a growing demand for tools to address the specific needs of 3D data stewardship. Central to this is the identification and creation of metadata models to describe these digital surrogates. Just as importantly, such metadata should describe the ‘raw’ source data from which 3D models are derived and should document the technical processes going into data collection and model creation. There are many institutions and organizations actively working in this area and we are publishing the Smithsonian’s internal 3D metadata model to add another data point to this critical discussion. We hope this will fuel further discussions within the community, helping everyone identify critical elements which can be standardized to facilitate robust data sharing, and pushing towards the creation of an interoperable metadata model.

Background
The Smithsonian is scaling up 3D digitization efforts across the institution, moving from low volume, high complexity projects towards high throughput production work aimed at capturing entire collections of objects as 3D models. For these activities, it is important to ensure that the collected data and resulting models are robust, reusable, and easily accessible. To meet these needs and to facilitate digitization activities, robust IT infrastructure that enables the preservation, management, and delivery of 3D data is essential. To this end, the Digitization Program Office (DPO) is developing a 3D data-centric suite of open source tools including a content management system, automated processing service, and web delivery platform. Underlying these tools is a metadata model which describes and brings transparency to the complex relationships between raw 3D capture data and the resulting models, ultimately making this data more durable and reusable.

Design Considerations
Before the first line of code was written for the tools mentioned above, the basic metadata model underlying them was well under development by an interdisciplinary working group of practitioners, informaticians, librarians, and archivists from across the Smithsonian Institution.  The group focused primarily on modeling the metadata needed to fully document a 3D capture event, and specifically focused on photogrammetry capture. Photogrammetry was chosen as a test case because it’s an accessible, non-proprietary technology and the ‘raw’ data, image files are also non-proprietary and and have existing best practices for preservation. Additionally, there is a high amount of complexity around how a photogrammetry project can be executed compared to other 3D capture methods. The hope is that, by addressing photogrammetry first, this model will already account for many “edge cases” in capture techniques and will be straightforward to extend to other capture types.

What’s Next? 
As there are no accepted metadata standards for 3D capture and processing information, the working group opted to designed much of this model from scratch, with the intent of swapping out terms from existing related schemas as the model matures.  The current version is an internal draft and it continues to evolve. We need your input to help us identify those existing terms, find areas that need improvement, and extend our work into an adaptable tool that can be used by others working with 3D data. Send us your feedback at OCIO-DPO@si.edu, or on one of our social media platforms.

Overview - Overview of 3D record types and metadata element attributes
Diagram - Diagram of record contents and relations between record types
Spreadsheet - Matrix of all 3D metadata elements
Visuals - A quick introduction to a few of the key concepts and records