TCT: A Cross-supervised Learning Method for Multimodal Sequence Representation.
Wubo LiWei ZouXiangang LiPublished in: CoRR (2019)
Keyphrases
- detection method
- supervised learning
- high precision
- segmentation method
- feature representation
- high accuracy
- probabilistic model
- pairwise
- clustering method
- computational cost
- experimental evaluation
- cost function
- prior knowledge
- fixed length
- representation scheme
- cross validation
- similarity measure
- generation method
- support vector machine svm
- unsupervised learning
- text classification
- input data
- edge detection
- image registration
- semi supervised
- active learning