Separability and refinement of hierarchical semantic video labels and their ground truth.
John R. KenderPublished in: ICME (2008)
Keyphrases
- ground truth
- semantic video
- classifier training
- image classification
- inter related
- multi label
- video event
- video search
- multi class
- video classification
- pairwise
- image annotation
- training data
- video collections
- semantic concepts
- surveillance videos
- video content
- visual concepts
- computer vision
- class labels
- wordnet
- object detection
- feature extraction