Training DCNN by Combining Max-Margin, Max-Correlation Objectives, and Correntropy Loss for Multilabel Image Classification.
Weiwei ShiYihong GongXiaoyu TaoNanning ZhengPublished in: IEEE Trans. Neural Networks Learn. Syst. (2018)
Keyphrases
- image classification
- multi label
- max margin
- classifier training
- multi label classification
- structured prediction
- visual features
- binary classification
- image annotation
- image features
- image representation
- inter related
- visual words
- multi class
- multiple labels
- feature extraction
- max margin markov
- graph cuts
- structured learning
- training set
- text categorization
- class labels
- learning tasks
- maximum margin
- training samples
- conditional random fields
- visual concepts
- multi task learning
- learning algorithm
- classification algorithm
- text classification
- supervised learning
- image processing
- feature selection