TOM-Net predicts an object mask, attenuation mask, and refractive flow field for transparent objects from one image via a multi-scale encoder-decoder plus residual refinement, trained on 178K synthetic images.
In: SIGGRAPH (1999) 2, 3, 4
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Learning Transparent Object Matting
TOM-Net predicts an object mask, attenuation mask, and refractive flow field for transparent objects from one image via a multi-scale encoder-decoder plus residual refinement, trained on 178K synthetic images.