A three-stage head-to-tail training recipe with optional unlabeled pseudo-labels achieves state-of-the-art long-tailed detection on LVIS v1 without external image-level supervision.
Cascade R-CNN: Delv- ing into High Quality Object Detection
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
2024 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
SimLTD: Simple Supervised and Semi-Supervised Long-Tailed Object Detection
A three-stage head-to-tail training recipe with optional unlabeled pseudo-labels achieves state-of-the-art long-tailed detection on LVIS v1 without external image-level supervision.