Deep Visual-Semantic Alignments for Generating Image Descriptions.
Andrej KarpathyLi Fei-FeiPublished in: CoRR (2014)
Keyphrases
- input image
- image features
- visual concepts
- multiscale
- image content
- single image
- low level
- image data
- visual perception
- image representation
- image description
- high level
- visual data
- visual appearance
- image classification
- image retrieval
- semantic labels
- web images
- textual descriptions
- visual cues
- image analysis
- image regions
- segmentation method
- semantic space
- semantic information
- semantically meaningful
- visual similarity
- similarity measure
- semantic gap
- human visual
- auto annotation
- edge detection
- natural language
- high level semantics
- semantic description
- semantic context
- web service descriptions
- semantic content
- pixel values
- image collections
- visual information
- spatial information
- keypoints
- visual features
- color images