Linear dimensionality reduction via a heteroscedastic extension of LDA: the Chernoff criterion

By Marco Loog, Robert P. W. Duin. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2004

322 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Subclass discriminant analysis (2006, methodology): “Heteroscedastic LDA [ 17 ] (HLDA) is such a method, where the authors use the Chernoff distance to estimate class similarity based on both means and covariances; i.e., is redened as”
Geometric Mean for Subspace Selection (2009, methodology): “4A C OMPARATIVE STUDY USING SYNTHETIC DATA In this section, we compare the effectiveness of MGMD with FLDA [13], HDA [21], aPAC [25], weighted FLDA (WFLDA) [13], FS-FLDA [23], HFLDA [ 26 ], ODA [7], and MODA [7] in handling three typical problems (heteroscedastic problem, multimodal problem, and cla”
Principal Component Analysis Based on L1-Norm Maximization (2008, methodology): “Th se datasets have been used in many studies [11] [12] [13].”
Automatic speech recognition and speech variability: A review (2007, methodology): “An exception is the method in [177], which uses the Chernoff distance to measure between-class distances and leads to a straight forward solution for A.”
Transferable Joint Attribute-Identity Deep Learning for Unsupervised Person Re-identification (2018, methodology): “mains are non-overlapping therefore presents a more challenging open-set recognition problem, as compared to the closed-set assumption made by most existing transfer learning models [35]; (3) The joint exploitation of attribute and identity labels gives rise to the heteroscedasticity (a mixture of d”