A causality-aware self-supervised pipeline reconstructs 3D non-rigid clothing from single images by embedding a structural causal map and two EM loops to disentangle camera, shape, texture, and illumination variables.
Dice: Domain-attack invariant causal learning for improved data privacy protection and adversarial robust- ness
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2022 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
3D Magic Mirror: Clothing Reconstruction from a Single Image via a Causal Perspective
A causality-aware self-supervised pipeline reconstructs 3D non-rigid clothing from single images by embedding a structural causal map and two EM loops to disentangle camera, shape, texture, and illumination variables.