Rethinking Feature Distribution for Loss Functions in Image Classification.
Weitao WanYuanyi ZhongTianpeng LiJiansheng ChenPublished in: CoRR (2018)
Keyphrases
- loss function
- image classification
- image features
- pairwise
- loss minimization
- logistic regression
- squared error
- learning to rank
- support vector
- feature extraction
- multi label
- update rules
- hinge loss
- risk minimization
- reproducing kernel hilbert space
- data distribution
- feature vectors
- probability distribution
- convex loss functions
- precision and recall
- object recognition