ELSA3D: Elastic Semantic Anchoring for Unified 3D Understanding and Generation

By Tianjiao Yu · Paper · cs.CV

Unified 3D foundation models aspire to generate 3D assets and reason about them in language within a single backbone, but their text-3D interaction remains largely implicit. Existing methods concatenate text and 3D tokens into a flat sequence and rely on self-attention, collapsin

Cs.cv

View original

HomeResourceLoading…