SepFormer uses a coarse-to-fine transformer decoder to regress table row and column separators in one shot, reaching 25.6 FPS with accuracy comparable to state-of-the-art TSR methods.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
SepFormer: Coarse-to-fine Separator Regression Network for Table Structure Recognition
SepFormer uses a coarse-to-fine transformer decoder to regress table row and column separators in one shot, reaching 25.6 FPS with accuracy comparable to state-of-the-art TSR methods.