Enhancing LiDAR Semantic Segmentation Robustness in Adverse Weather through Domain Adaptation and Generalization
UniMix, a universal method that enhances the adaptability and generalizability of LiDAR semantic segmentation models to unseen adverse-weather scenes.
Segment Anything in Lidar: Distilling Vision Foundation Models for Zero-Shot Panoptic Segmentation
SAL, a method for Zero-Shot Lidar Panoptic Segmentation, utilizes a pseudo-label engine that distills vision foundation models to Lidar and a zero-shot model trained via self-supervision. This allows for segmenting and classifying any object in a Lidar scan without manual supervision.
LiDARFormer: A Unified Transformer-based Multi-task Network for LiDAR Perception
새로운 LiDAR 멀티태스크 네트워크인 LiDARFormer은 효율적인 LiDAR 지각을 위한 혁신적인 패러다임을 제시합니다.
LiDARFormer: A Unified Transformer-based Multi-task Network for LiDAR Perception
新しいLiDARマルチタスクネットワークのコアメッセージは、トランスフォーマーを基盤としたパラダイムであり、検出とセグメンテーションのタスクを効果的に統合し、競争力のある成績を達成することです。
LiDARFormer: Transformer-based Multi-task Network for LiDAR Perception
The author introduces LiDARFormer, a novel multi-task network based on transformers, to enhance LiDAR perception tasks by leveraging cross-space and cross-task attention.
© 2024 by Linnk AI