PointCSP achieves better semantic consistency and performance in point cloud self-supervised learning by propagating semantics across batch samples via state-space models and stabilizing transfer with asymmetric distillation.
Emerg- ing properties in self-supervised vision transformers
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.CV 3verdicts
UNVERDICTED 3representative citing papers
SENSE is the first stereo open-vocabulary semantic segmentation method that uses vision-language models and stereo geometry to achieve better phrase-grounded segmentation and generalization on benchmarks like Cityscapes and KITTI.
An unsupervised monocular road segmentation method generates initial labels from horizon and quadrilateral geometric priors then refines them via temporal feature tracking and mutual information to reach 0.86 IoU on Cityscapes.
citing papers explorer
-
PointCSP: Cross-Sample Semantic Propagation and Stability Preservation in Self-Supervised Point Cloud Learning
PointCSP achieves better semantic consistency and performance in point cloud self-supervised learning by propagating semantics across batch samples via state-space models and stabilizing transfer with asymmetric distillation.
-
SENSE: Stereo OpEN Vocabulary SEmantic Segmentation
SENSE is the first stereo open-vocabulary semantic segmentation method that uses vision-language models and stereo geometry to achieve better phrase-grounded segmentation and generalization on benchmarks like Cityscapes and KITTI.
-
Unsupervised Monocular Road Segmentation for Autonomous Driving via Scene Geometry
An unsupervised monocular road segmentation method generates initial labels from horizon and quadrilateral geometric priors then refines them via temporal feature tracking and mutual information to reach 0.86 IoU on Cityscapes.