Evaluating Multimodal Representations on Visual Semantic Textual Similarity.
Oier Lopez de LacalleAnder SalaberriaAitor SoroaGorka AzkuneEneko AgirrePublished in: CoRR (2020)
Keyphrases
- visual representations
- semantic similarity
- semantic representations
- natural language
- semantic labels
- visual content
- visual information
- similarity measure
- visual features
- sentence similarity
- semantic context
- visual similarity
- semantic content
- semantic representation
- higher level
- semantic similarity measure
- cross modal
- visual patterns
- semantically relevant
- multimodal information
- mid level
- semantic concepts
- multimedia
- multi modal
- low level
- visual cues
- natural language understanding
- semantic information
- semantic web
- high level
- metadata
- domain specific
- semantic space
- distance measure
- visual concepts
- intermediate representations
- keywords
- image retrieval
- word similarity
- semantic distance
- multiple modalities
- wordnet
- visualization tools
- euclidean distance
- domain ontology
- audio visual
- distance function
- co occurrence
- single modality
- similarity relations
- object recognition
- textual information
- visual and textual features