Moment invariants for 3D flow fields via normalization

We generalize the framework of moments and introduce a definition of invariants for three-dimensional vector fields. To do so, we use the method of moment normalization that has been shown to be useful in the two dimensions. Using invariant moments, we show how to search for patterns in these fields...

Full description

Saved in:
Bibliographic Details
Published inIEEE Pacific Visualization Symposium pp. 9 - 16
Main Authors Bujack, Roxana, Kasten, Jens, Hotz, Ingrid, Scheuermann, Gerik, Hitzer, Eckhard
Format Conference Proceeding
LanguageEnglish
Japanese
Published IEEE 01.04.2015
Subjects
Covariance matrices
I.4.7 [Image Processing and Computer Vision]: Feature Measurement -Moments
I.5.2 [Pattern Recognition]: Design Methodology -Classifier design and evaluation
Matrices
Symmetric matrices
Tensile stress
Three-dimensional displays
Transforms
Online AccessGet full text
ISSN2165-8765
DOI10.1109/PACIFICVIS.2015.7156350

Cover

More Information
Summary:We generalize the framework of moments and introduce a definition of invariants for three-dimensional vector fields. To do so, we use the method of moment normalization that has been shown to be useful in the two dimensions. Using invariant moments, we show how to search for patterns in these fields independent from their position, orientation and scale. From the first order vector moment tensor, we construct a complete and independent set of descriptors. We test the invariants in queries on synthetic and real world flow fields.
ISSN:2165-8765
DOI:10.1109/PACIFICVIS.2015.7156350