Bridging the Visual Semantic Gap in VLN via Semantically Richer Instructions.
Joaquín OssandónBenjamín EarleÁlvaro SotoPublished in: ECCV (37) (2022)
Keyphrases
- semantic gap
- visual features
- low level
- semantic information
- semantic content
- low level features
- bridge the semantic gap
- visual content
- visual information
- high level semantics
- video retrieval
- visual concepts
- image content
- semantic concepts
- high level
- low level visual features
- image classification
- image annotation
- higher level
- image retrieval
- visual data
- domain knowledge
- multimedia databases
- semantic similarity
- image collections
- relational databases
- contextual information
- wordnet
- multiscale
- multi modal