Rethinking semantic-visual alignment in zero-shot object detection via a softplus margin focal loss.
Qianzhong LiYujia ZhangShiying SunXiaoguang ZhaoKang LiMin TanPublished in: Neurocomputing (2021)
Keyphrases
- object detection
- object categories
- visual concepts
- semantic content
- high level
- face detection
- computer vision
- object class
- visual context
- multi class
- low level
- category level
- visual features
- visual information
- natural language
- semantic information
- ontology matching
- semantic network
- scene understanding
- hinge loss
- contextual cues
- object classes
- semantic concepts
- domain specific
- semantic web
- low level features
- machine learning
- semantically relevant
- face recognition
- semantically related
- semantic similarity
- semantic search
- sequence alignment
- image alignment
- pedestrian detection
- maximum margin
- visual perception
- decision boundary