Capturing Localized Image Artifacts through a CNN-based Hyper-image Representation.
Parag Shridhar ChandakkarBaoxin LiPublished in: CoRR (2017)
Keyphrases
- image representation
- image content
- image classification
- multiscale
- image features
- image retrieval
- object recognition
- bag of words
- sparse coding
- image description
- quadtree
- low level features
- bag of features
- visual words
- feature space
- visual content
- scene recognition
- bag of visual words
- sparse representation
- representation scheme
- compressive sensing
- region segmentation
- scene categorization
- scene classification
- spatial information
- image collections
- edge detection
- spatial pyramid matching