Feature Selection for Unsupervised Anomaly Detection and Localization Using Synthetic Defects.
Lars HecklerRebecca KönigPublished in: VISIGRAPP (3): VISAPP (2024)
Keyphrases
- feature selection
- unsupervised feature selection
- unsupervised learning
- pointwise mutual information
- machine learning
- mutual information
- supervised and unsupervised learning
- text categorization
- model selection
- dimensionality reduction
- supervised learning
- semi supervised
- microarray data
- feature selection algorithms
- selected features
- unsupervised manner
- irrelevant features
- informative features
- text classification
- multi class
- support vector
- support vector machine
- data driven
- feature extraction
- feature space
- small sample
- discriminative features
- classification accuracy
- feature selection and classification
- information gain
- completely unsupervised
- real images are presented
- method for feature selection
- high dimensionality
- supervised learning tasks
- defect classification
- defect detection
- supervised classification
- classification models
- data sets
- knn
- image segmentation
- data mining
- real world