RT-NeRV discretizes residual features in hybrid NeRV into compact tokens via a residual tokenizer and codebook learning strategy, improving detail preservation and bitrate trade-offs in video regression and restoration tasks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
RT-NeRV: Rethinking Hybrid Neural Representations for Video via Residual Tokenization
RT-NeRV discretizes residual features in hybrid NeRV into compact tokens via a residual tokenizer and codebook learning strategy, improving detail preservation and bitrate trade-offs in video regression and restoration tasks.