The impact of Compositionality in Zero-shot Multi-label action recognition for Object-based tasks.
Carmela CalabreseStefano BertiGiulia PasqualeLorenzo NatalePublished in: CoRR (2024)
Keyphrases
- action recognition
- multi label
- multi label classification
- text categorization
- human actions
- graph cuts
- binary classification
- image classification
- image annotation
- bag of words
- max margin
- computer vision
- text classification
- bag of features
- learning tasks
- automatic image annotation
- recognizing human actions
- class labels
- human activities
- multi modal
- image processing
- transfer learning
- markov random field
- image features
- spatio temporal
- training data