TY - GEN
T1 - Generalized dictionary learning for symmetric positive definite matrices with application to nearest neighbor retrieval
AU - Sra, Suvrit
AU - Cherian, Anoop
PY - 2011
Y1 - 2011
N2 - We introduce Generalized Dictionary Learning (GDL), a simple but practical framework for learning dictionaries over the manifold of positive definite matrices. We illustrate GDL by applying it to Nearest Neighbor (NN) retrieval, a task of fundamental importance in disciplines such as machine learning and computer vision. GDL distinguishes itself from traditional dictionary learning approaches by explicitly taking into account the manifold structure of the data. In particular, GDL allows performing "sparse coding" of positive definite matrices, which enables better NN retrieval. Experiments on several covariance matrix datasets show that GDL achieves performance rivaling state-of-the-art techniques.
AB - We introduce Generalized Dictionary Learning (GDL), a simple but practical framework for learning dictionaries over the manifold of positive definite matrices. We illustrate GDL by applying it to Nearest Neighbor (NN) retrieval, a task of fundamental importance in disciplines such as machine learning and computer vision. GDL distinguishes itself from traditional dictionary learning approaches by explicitly taking into account the manifold structure of the data. In particular, GDL allows performing "sparse coding" of positive definite matrices, which enables better NN retrieval. Experiments on several covariance matrix datasets show that GDL achieves performance rivaling state-of-the-art techniques.
UR - https://www.scopus.com/pages/publications/80052399488
U2 - 10.1007/978-3-642-23808-6_21
DO - 10.1007/978-3-642-23808-6_21
M3 - Conference contribution
AN - SCOPUS:80052399488
SN - 9783642238079
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 318
EP - 332
BT - Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2011, Proceedings
PB - Springer Verlag
ER -