Geometries and interpolations for symmetric positive definite matrices
Research output: Chapter in Book/Report/Conference proceeding › Book chapter › Research › peer-review
Standard
Geometries and interpolations for symmetric positive definite matrices. / Feragen, Aasa; Fuster, Andrea.
Modeling, analysis, and visualization of anisotropy. ed. / Thomas Schultz; Evren Özarslan; Ingrid Hotz. Springer, 2017. p. 85-113 (Mathematics and Visualization).Research output: Chapter in Book/Report/Conference proceeding › Book chapter › Research › peer-review
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - CHAP
T1 - Geometries and interpolations for symmetric positive definite matrices
AU - Feragen, Aasa
AU - Fuster, Andrea
PY - 2017
Y1 - 2017
N2 - In this survey we review classical and recently proposed Riemannian metrics and interpolation schemes on the space of symmetric positive definite (SPD) matrices. We perform simulations that illustrate the problem of tensor fattening not only in the usually avoided Frobenius metric, but also in other classical metrics on SPD matrices such as the Wasserstein metric, the affine invariant/Fisher Rao metric, and the log Euclidean metric. For comparison, we perform the same simulations on several recently proposed frameworks for SPD matrices that decompose tensors into shape and orientation. In light of the simulation results, we discuss the mathematical and qualitative properties of these new metrics in comparison with the classical ones. Finally, we explore the nonlinear variation of properties such as shape and scale throughout principal geodesics in different metrics, which affects the visualization of scale and shape variation in tensorial data. With the paper, we will release a software package with Matlab scripts for computing the interpolations and statistics used for the experiments in the paper (Code is available at https://sites.google.com/site/aasaferagen/home/software).
AB - In this survey we review classical and recently proposed Riemannian metrics and interpolation schemes on the space of symmetric positive definite (SPD) matrices. We perform simulations that illustrate the problem of tensor fattening not only in the usually avoided Frobenius metric, but also in other classical metrics on SPD matrices such as the Wasserstein metric, the affine invariant/Fisher Rao metric, and the log Euclidean metric. For comparison, we perform the same simulations on several recently proposed frameworks for SPD matrices that decompose tensors into shape and orientation. In light of the simulation results, we discuss the mathematical and qualitative properties of these new metrics in comparison with the classical ones. Finally, we explore the nonlinear variation of properties such as shape and scale throughout principal geodesics in different metrics, which affects the visualization of scale and shape variation in tensorial data. With the paper, we will release a software package with Matlab scripts for computing the interpolations and statistics used for the experiments in the paper (Code is available at https://sites.google.com/site/aasaferagen/home/software).
UR - http://www.scopus.com/inward/record.url?scp=85032013458&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-61358-1_5
DO - 10.1007/978-3-319-61358-1_5
M3 - Book chapter
AN - SCOPUS:85032013458
SN - 978-3-319-61357-4
T3 - Mathematics and Visualization
SP - 85
EP - 113
BT - Modeling, analysis, and visualization of anisotropy
A2 - Schultz, Thomas
A2 - Özarslan, Evren
A2 - Hotz, Ingrid
PB - Springer
ER -
ID: 188486137