OVIS: Open-Vocabulary Visual Instance Search via Visual-Semantic Aligned Representation Learning.
Sheng LiuKevin LinLijuan WangJunsong YuanZicheng LiuPublished in: AAAI (2022)
Keyphrases
- visual learning
- visual representation
- perceptual information
- visual features
- active learning
- visual information
- learning systems
- domain independent
- auto annotation
- semantic search
- search algorithm
- prior knowledge
- search space
- visual properties
- reinforcement learning
- visual search
- visual representations
- video search
- learning algorithm
- domain specific
- search engine
- semantic space
- visual processing
- keywords
- visual patterns
- semantic content
- learning tasks
- natural language
- neural network
- domain ontology
- background knowledge