A self-supervised training recipe that clusters features into semantic, instance, and hierarchical groups, then uses the grouped representations to improve downstream object detection.
In: Proceedings of the 36th Conference on Neural Inform ation Processing Sys- tems (NeurIPS) (2022) 12 O
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
-
Self-supervised structured object representation learning
A self-supervised training recipe that clusters features into semantic, instance, and hierarchical groups, then uses the grouped representations to improve downstream object detection.