Riemannian Variance Filtering: An Independent Filtering Scheme for Statistical Tests on Manifold-Valued Data

Ligang Zheng, Hyun Woo Kim, Nagesh Adluru, Michael A. Newton, Vikas Singh

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Performing large scale hypothesis testing on brain imaging data to identify group-wise differences (e.g., between healthy and diseased subjects) typically leads to a large number of tests (one per voxel). Multiple testing adjustment (or correction) is necessary to control false positives, which may lead to lower detection power in detecting true positives. Motivated by the use of socalled 'independent filtering' techniques in statistics (for genomics applications), this paper investigates the use of independent filtering for manifold-valued data (e.g., Diffusion Tensor Imaging, Cauchy Deformation Tensors) which are broadly used in neuroimaging studies. Inspired by the concept of variance of a Riemannian Gaussian distribution, a type of non-specific data-dependent Riemannian variance filter is proposed. In practice, the filter will select a subset of the full set of voxels for performing the statistical test, leading to a more appropriate multiple testing correction. Our experiments on synthetic/simulated manifoldvalued data show that the detection power is improved when the statistical tests are performed on the voxel locations that 'pass' the filter. Given the broadening scope of applications where manifold-valued data are utilized, the scheme can serve as a general feature selection scheme.

Original languageEnglish
Title of host publicationProceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017
PublisherIEEE Computer Society
Pages699-708
Number of pages10
ISBN (Electronic)9781538607336
DOIs
Publication statusPublished - 2017 Aug 22
Externally publishedYes
Event30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017 - Honolulu, United States
Duration: 2017 Jul 212017 Jul 26

Publication series

NameIEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops
Volume2017-July
ISSN (Print)2160-7508
ISSN (Electronic)2160-7516

Conference

Conference30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017
CountryUnited States
CityHonolulu
Period17/7/2117/7/26

Fingerprint

Statistical tests
Testing
Diffusion tensor imaging
Neuroimaging
Gaussian distribution
Tensors
Feature extraction
Brain
Statistics
Imaging techniques
Experiments

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Electrical and Electronic Engineering

Cite this

Zheng, L., Kim, H. W., Adluru, N., Newton, M. A., & Singh, V. (2017). Riemannian Variance Filtering: An Independent Filtering Scheme for Statistical Tests on Manifold-Valued Data. In Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017 (pp. 699-708). [8014833] (IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops; Vol. 2017-July). IEEE Computer Society. https://doi.org/10.1109/CVPRW.2017.99

Riemannian Variance Filtering : An Independent Filtering Scheme for Statistical Tests on Manifold-Valued Data. / Zheng, Ligang; Kim, Hyun Woo; Adluru, Nagesh; Newton, Michael A.; Singh, Vikas.

Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017. IEEE Computer Society, 2017. p. 699-708 8014833 (IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops; Vol. 2017-July).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Zheng, L, Kim, HW, Adluru, N, Newton, MA & Singh, V 2017, Riemannian Variance Filtering: An Independent Filtering Scheme for Statistical Tests on Manifold-Valued Data. in Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017., 8014833, IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, vol. 2017-July, IEEE Computer Society, pp. 699-708, 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017, Honolulu, United States, 17/7/21. https://doi.org/10.1109/CVPRW.2017.99
Zheng L, Kim HW, Adluru N, Newton MA, Singh V. Riemannian Variance Filtering: An Independent Filtering Scheme for Statistical Tests on Manifold-Valued Data. In Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017. IEEE Computer Society. 2017. p. 699-708. 8014833. (IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops). https://doi.org/10.1109/CVPRW.2017.99
Zheng, Ligang ; Kim, Hyun Woo ; Adluru, Nagesh ; Newton, Michael A. ; Singh, Vikas. / Riemannian Variance Filtering : An Independent Filtering Scheme for Statistical Tests on Manifold-Valued Data. Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017. IEEE Computer Society, 2017. pp. 699-708 (IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops).
@inproceedings{db941b14e2614e39ab01c8a6869ac59f,
title = "Riemannian Variance Filtering: An Independent Filtering Scheme for Statistical Tests on Manifold-Valued Data",
abstract = "Performing large scale hypothesis testing on brain imaging data to identify group-wise differences (e.g., between healthy and diseased subjects) typically leads to a large number of tests (one per voxel). Multiple testing adjustment (or correction) is necessary to control false positives, which may lead to lower detection power in detecting true positives. Motivated by the use of socalled 'independent filtering' techniques in statistics (for genomics applications), this paper investigates the use of independent filtering for manifold-valued data (e.g., Diffusion Tensor Imaging, Cauchy Deformation Tensors) which are broadly used in neuroimaging studies. Inspired by the concept of variance of a Riemannian Gaussian distribution, a type of non-specific data-dependent Riemannian variance filter is proposed. In practice, the filter will select a subset of the full set of voxels for performing the statistical test, leading to a more appropriate multiple testing correction. Our experiments on synthetic/simulated manifoldvalued data show that the detection power is improved when the statistical tests are performed on the voxel locations that 'pass' the filter. Given the broadening scope of applications where manifold-valued data are utilized, the scheme can serve as a general feature selection scheme.",
author = "Ligang Zheng and Kim, {Hyun Woo} and Nagesh Adluru and Newton, {Michael A.} and Vikas Singh",
year = "2017",
month = "8",
day = "22",
doi = "10.1109/CVPRW.2017.99",
language = "English",
series = "IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops",
publisher = "IEEE Computer Society",
pages = "699--708",
booktitle = "Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017",

}

TY - GEN

T1 - Riemannian Variance Filtering

T2 - An Independent Filtering Scheme for Statistical Tests on Manifold-Valued Data

AU - Zheng, Ligang

AU - Kim, Hyun Woo

AU - Adluru, Nagesh

AU - Newton, Michael A.

AU - Singh, Vikas

PY - 2017/8/22

Y1 - 2017/8/22

N2 - Performing large scale hypothesis testing on brain imaging data to identify group-wise differences (e.g., between healthy and diseased subjects) typically leads to a large number of tests (one per voxel). Multiple testing adjustment (or correction) is necessary to control false positives, which may lead to lower detection power in detecting true positives. Motivated by the use of socalled 'independent filtering' techniques in statistics (for genomics applications), this paper investigates the use of independent filtering for manifold-valued data (e.g., Diffusion Tensor Imaging, Cauchy Deformation Tensors) which are broadly used in neuroimaging studies. Inspired by the concept of variance of a Riemannian Gaussian distribution, a type of non-specific data-dependent Riemannian variance filter is proposed. In practice, the filter will select a subset of the full set of voxels for performing the statistical test, leading to a more appropriate multiple testing correction. Our experiments on synthetic/simulated manifoldvalued data show that the detection power is improved when the statistical tests are performed on the voxel locations that 'pass' the filter. Given the broadening scope of applications where manifold-valued data are utilized, the scheme can serve as a general feature selection scheme.

AB - Performing large scale hypothesis testing on brain imaging data to identify group-wise differences (e.g., between healthy and diseased subjects) typically leads to a large number of tests (one per voxel). Multiple testing adjustment (or correction) is necessary to control false positives, which may lead to lower detection power in detecting true positives. Motivated by the use of socalled 'independent filtering' techniques in statistics (for genomics applications), this paper investigates the use of independent filtering for manifold-valued data (e.g., Diffusion Tensor Imaging, Cauchy Deformation Tensors) which are broadly used in neuroimaging studies. Inspired by the concept of variance of a Riemannian Gaussian distribution, a type of non-specific data-dependent Riemannian variance filter is proposed. In practice, the filter will select a subset of the full set of voxels for performing the statistical test, leading to a more appropriate multiple testing correction. Our experiments on synthetic/simulated manifoldvalued data show that the detection power is improved when the statistical tests are performed on the voxel locations that 'pass' the filter. Given the broadening scope of applications where manifold-valued data are utilized, the scheme can serve as a general feature selection scheme.

UR - http://www.scopus.com/inward/record.url?scp=85030217705&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85030217705&partnerID=8YFLogxK

U2 - 10.1109/CVPRW.2017.99

DO - 10.1109/CVPRW.2017.99

M3 - Conference contribution

AN - SCOPUS:85030217705

T3 - IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops

SP - 699

EP - 708

BT - Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017

PB - IEEE Computer Society

ER -