Proposes the first unified incomplete video-language model that processes missing modalities and serves as a plug-and-play module to boost existing VLMs on multi-modal tasks.
InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 14773–14783
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
-
Towards Unified Vision-Language Models with Incomplete Multi-Modal Inputs
Proposes the first unified incomplete video-language model that processes missing modalities and serves as a plug-and-play module to boost existing VLMs on multi-modal tasks.