Who Generated This 3D Asset? Learning Source Attribution for Generative 3D Models

In three linesFirst systematic study of source attribution for generated 3D assets. Researchers build a benchmark covering 22 3D generators and propose a hierarchical multi-view multi-modal Transformer detecting fingerprints (cross-view inconsistencies, geometric artifacts, frequency-domain signatures). Results: 97.22% accuracy under full supervision, 77.17% with only 1% training data.

Summary generated by Claude — human-verified