MARMOT shows that a masked autoencoder pretrained on a large synthetic transient dataset can complete heavily masked non-line-of-sight transients and transfer to downstream NLOS tasks.
Pretraining a foundation model for generalizable fluorescence microscopy-based image restoration,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
support 1representative citing papers
citing papers explorer
-
MARMOT: Masked Autoencoder for Modeling Transient Imaging
MARMOT shows that a masked autoencoder pretrained on a large synthetic transient dataset can complete heavily masked non-line-of-sight transients and transfer to downstream NLOS tasks.