Fixed-point iterative computation of Gaussian barycenters for some dissimilarity measures

Date

2022-01-01

Author

D'ortenzio, Alessandro
Manes, Costanzo
Orguner, Umut

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

1
views

0
downloads

Cite This

In practical contexts like sensor fusion or computer vision, it is not unusual to deal with a large number of Gaussian densities that encode the available information. In such cases, if the computational capabilities are limited, a data compression is required, often done by finding the barycenter of the set of Gaussians. However, such computation strongly depends on the chosen loss function (dissimilarity measure) to be minimized, and most often it must be performed by means of numerical methods, since the barycenter can rarely be computed analytically. Some constraints, like the covariance matrix symmetry and positive definiteness can make nontrivial the numerical computation of the Gaussian barycenter. In this work, a set of Fixed-Point Iteration algorithms are presented in order to allow for the agile computation of Gaussian barycenters according to several dissimilarity measures.

Subject Keywords

barycenters, data compression, fixed-point iterations, Gaussian densities, signal processing

URI

https://hdl.handle.net/11511/106613

Collections

Department of Electrical and Electronics Engineering, Conference / Seminar