CHARM is a modality-agnostic segmentation framework that implicitly aligns different sensor inputs while preserving each modality's specific strengths through a mutual perception unit and a dual-path training strategy.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
CHARM: Collaborative Harmonization across Arbitrary Modalities for Modality-agnostic Semantic Segmentation
CHARM is a modality-agnostic segmentation framework that implicitly aligns different sensor inputs while preserving each modality's specific strengths through a mutual perception unit and a dual-path training strategy.