Cyclic Annealing Training Convolutional Neural Networks for Image Classification with Noisy Labels.
Jiawei LiTao DaiQingtao TangYeli XingShu-Tao XiaPublished in: ICIP (2018)
Keyphrases
- image classification
- convolutional neural networks
- multi label
- training examples
- classifier training
- training set
- bag of words
- visual features
- training process
- simulated annealing
- supervised learning
- image representation
- noisy data
- training data
- feature extraction
- convolutional network
- ground truth labels
- labelled data
- training samples
- test set
- training algorithm
- training phase
- class specific
- image features
- data sets
- online learning
- hidden markov models
- remotely sensed data
- image processing
- annotation effort
- learning algorithm