Unsupervised training scheme with non-stereo data for empirical feature vector compensation.
Luis BueraAntonio MiguelAlfonso OrtegaEduardo LleidaRichard M. SternPublished in: INTERSPEECH (2009)
Keyphrases
- data sets
- data collection
- input data
- feature vectors
- prior knowledge
- synthetic data
- data structure
- raw data
- small number
- database
- training data
- data analysis
- high quality
- training set
- data points
- feature extraction
- data processing
- data mining
- original data
- xml documents
- feature space
- image features
- dimensionality reduction
- computer vision
- training samples
- high dimensional data
- empirical data