CrossView Suite supplies a 1.6M-sample dataset, scene-disjoint benchmark, and explicit-alignment framework to advance MLLMs from single-view perception to cross-view spatial intelligence.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
CrossView Suite: Harnessing Cross-view Spatial Intelligence of MLLMs with Dataset, Model and Benchmark
CrossView Suite supplies a 1.6M-sample dataset, scene-disjoint benchmark, and explicit-alignment framework to advance MLLMs from single-view perception to cross-view spatial intelligence.