Docomomo 3D Platform: Open, Multi-Modal Models for Modern Architectural Heritage

Abstract

Since its foundation in 1990 Docomomo International has created a body of knowledge through their documentation “fiches” produced by the national working parties and coordinated by the International Specialist Committee (ISC) Registers. Although based on literature and other publications the fiches often do not contain accurate and complete plans and layouts and little spatial information. The Docomomo 3D Platform addresses this last gap by aggregating and publishing multi-modal spatial content of modern movement architecture, and is designed for clarity, interoperability, and public reuse. The platform curates four primary content types: (1) smart point-cloud models that embed semantic tags and condition notes; (2) 3D Gaussian Splatting (3DGS) reconstructions optimized for photoreal, real-time exploration; (3) HBIM models that encode construction logic, materials, and interventions for conservation planning; and (4) synchronized 2D assets (photographs and short videos) that contextualize each building’s design, setting, and use. A consistent metadata schema links these modalities at the building level, enabling users to pivot from a navigable 3D experience to archival stills, plans, and bibliographic notes. The next development phase introduces participatory pipelines: users will be able to upload image sets that the platform processes into community-generated 3DGS models, with automated quality checks and curator review. This opens a path for distributed recording of at-risk sites and creates a living repository that tracks change over time. By unifying documentation types and lowering the barrier to contribution, the Docomomo 3D Platform advances two goals: to improve public understanding of modern architecture and design, its materials, space, and everyday life, and to accelerate responsible dissemination for research, education, advocacy, and finally conservation. The paper outlines the data standards, ingestion workflow, and viewer UX, and presents case studies demonstrating how multi-modal models clarify significance, integrity, and conservation options for modern heritage.