The use of deep learning features in a hierarchical classifier learned with the minimization of a non-greedy loss function that delays gratification.
Zhibin LiaoGustavo CarneiroPublished in: ICIP (2015)
Keyphrases
- loss function
- deep learning
- unsupervised feature learning
- feature set
- support vector
- empirical risk
- feature space
- svm classifier
- deep belief networks
- feature vectors
- feature selection
- restricted boltzmann machine
- pairwise
- machine learning
- higher order
- training data
- co occurrence
- image features
- feature subset
- unsupervised learning
- training examples
- learning algorithm
- training samples
- learning strategies
- support vector machine
- feature extraction
- decision trees