Self-supervision for medical image classification: state-of-the-art performance with ~100 labeled training samples per class.
Maximilian NielsenLaura WenderothThilo SentkerRené WernerPublished in: CoRR (2023)
Keyphrases
- training samples
- test sample
- image classification
- supervised learning
- training data
- training set
- class specific
- feature space
- learning algorithm
- number of training samples
- sample set
- high dimensional
- active learning
- training examples
- hyperplane
- unlabeled samples
- bag of words
- image representation
- decision boundary
- positive training examples
- positive examples
- object classes
- data sets
- svm classifier
- sparse representation
- linear combination
- support vectors
- feature extraction
- decision function
- image processing
- model selection
- image features
- support vector machine
- machine learning
- database