M3D-GAN introduces modality subnets, a shared computing body, and a universal attention module to enable translation across text, image, and speech modalities and their internal domains.
• Image→Image: In this task, the source and target are images drawn from two different domains (e.g
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
M3D-GAN: Multi-Modal Multi-Domain Translation with Universal Attention
M3D-GAN introduces modality subnets, a shared computing body, and a universal attention module to enable translation across text, image, and speech modalities and their internal domains.