A multi-resolution vision-transformer model achieves fine-grained material selection in images at both texture and subtexture levels, trained on a new 800,000-image synthetic dataset.
In International Conference on Machine Learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.GR 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Fine-Grained Spatially Varying Material Selection in Images
A multi-resolution vision-transformer model achieves fine-grained material selection in images at both texture and subtexture levels, trained on a new 800,000-image synthetic dataset.