BiometricBlender: Ultra-high dimensional, multi-class synthetic data generator to imitate biometric feature space.
Marcell StippingerDávid HanákMarcell T. KurbuczGergely HanczárOlivér M. TörteliZoltán SomogyváriPublished in: CoRR (2022)
Keyphrases
- synthetic data
- multi class
- high dimensional
- feature space
- support vector machine
- feature selection
- feature extraction
- low dimensional
- high dimensionality
- dimensionality reduction
- input space
- real world
- high dimensional feature space
- data points
- feature vectors
- data sets
- high dimensional data
- nearest neighbor
- nearest neighbor search
- cost sensitive
- multi class classification
- binary classification
- training samples
- kernel function
- object detection
- classification accuracy
- kernel methods
- principal component analysis
- multi class object detection
- real image data
- multiple classes
- pairwise
- feature set
- multi task
- original data
- variable selection
- linear classifiers
- hyperplane
- binary classifiers
- learning algorithm
- error correcting output codes
- svm classifier