Harvesting Mid-level Visual Concepts from Large-Scale Internet Images.
Quannan LiJiajun WuZhuowen TuPublished in: CVPR (2013)
Keyphrases
- mid level
- visual concepts
- image collections
- low level
- low level features
- high level
- image content
- video content
- image annotation
- object categories
- semantic gap
- learning tasks
- semantic concepts
- visual information
- visual content
- event detection
- action recognition
- superpixels
- visual cues
- visual data
- image features
- positive examples
- image retrieval
- image classification
- markov random field
- image data
- object recognition
- semantic similarity
- image understanding
- video streams
- multi modal
- training data
- machine learning