MDQE: Mining Discriminative Query Embeddings to Segment Occluded Instances on Challenging Videos.
Minghan LiShuai LiWangmeng XiangLei ZhangPublished in: CoRR (2023)
Keyphrases
- video segments
- vector space
- database
- response time
- query processing
- video sequences
- category labels
- video frames
- query expansion
- user queries
- street view
- video dataset
- query evaluation
- data mining
- information retrieval
- video search
- relevance feedback
- text mining
- web mining
- knowledge discovery
- video clips
- video content
- manifold learning
- action classification
- video retrieval
- semi supervised
- pattern mining
- feature extraction
- keywords
- data structure
- action recognition
- data sources
- range queries
- low dimensional
- human actions
- database systems
- textual descriptions
- multimedia