CLIP-VIS: Adapting CLIP for Open-Vocabulary Video Instance Segmentation.
Wenqi ZhuJiale CaoJin XieShuangming YangYanwei PangPublished in: CoRR (2024)
Keyphrases
- video clips
- video database
- video frames
- video segmentation
- key frames
- video data
- video content
- video scene
- image segmentation
- video segments
- level set
- video streams
- video shots
- segmentation algorithm
- video sequences
- multiscale
- video retrieval
- region growing
- real time
- long video
- segmentation method
- shape prior
- image analysis
- fully automatic
- energy function
- space time
- edge detection
- fully unsupervised
- foreground background segmentation
- multimedia
- temporal information
- mean shift
- object segmentation
- medical images
- motion estimation
- feature vectors