LiDAL: Inter-frame Uncertainty Based Active Learning for 3D LiDAR Semantic Segmentation.
Zeyu HuXuyang BaiRunze ZhangXin WangGuangyuan SunHongbo FuChiew-Lan TaiPublished in: CoRR (2022)
Keyphrases
- inter frame
- semantic segmentation
- active learning
- motion estimation
- superpixels
- video sequences
- conditional random fields
- bit rate
- scene classification
- weakly supervised
- video frames
- object categories
- object classes
- quadtree
- supervised learning
- training examples
- semi supervised
- pixel level
- high resolution
- super resolution
- machine learning
- image set
- computer vision
- image sequences
- multiscale
- visual tracking
- optical flow
- image quality
- image registration