A comparative study of supervised learning algorithms for symmetric positive definite features

Ammar Mian; Elias Raninen; Esa Ollila

doi:10.23919/Eusipco47968.2020.9287531

A comparative study of supervised learning algorithms for symmetric positive definite features

Ammar Mian, Elias Raninen, Esa Ollila

Esa Ollila Group

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

143 Lataukset (Pure)

Abstrakti

In recent years, the use of Riemannian geometry has reportedly shown an increased performance for machine learning problems whose features lie in the symmetric positive definite (SPD) manifold. The present paper aims at reviewing several approaches based on this paradigm and provide a reproducible comparison of their output on a classic learning task of pedestrian detection. Notably, the robustness of these approaches to corrupted data will be assessed.

Alkuperäiskieli	Englanti
Otsikko	28th European Signal Processing Conference, EUSIPCO 2020 - Proceedings
Kustantaja	European Association For Signal and Image Processing
Sivut	950-954
Sivumäärä	5
ISBN (elektroninen)	9789082797053
DOI - pysyväislinkit	https://doi.org/10.23919/Eusipco47968.2020.9287531
Tila	Julkaistu - 2020
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisussa
Tapahtuma	European Signal Processing Conference - Amsterdam, Alankomaat Kesto: 24 elok. 2020 → 28 elok. 2020 Konferenssinumero: 28

Julkaisusarja

Nimi	European Signal Processing Conference
ISSN (painettu)	2219-5491
ISSN (elektroninen)	2076-1465

Conference

Conference	European Signal Processing Conference
Lyhennettä	EUSIPCO
Maa/Alue	Alankomaat
Kaupunki	Amsterdam
Ajanjakso	24/08/2020 → 28/08/2020

Pääsy asiakirjaan

10.23919/Eusipco47968.2020.9287531

Mian_Comparative_study_eusipcoFinal published version, 389 KB

https://www.eurasip.org/Proceedings/Eusipco/Eusipco2020/pdfs/0000950.pdf

Muut tiedostot ja linkit

Link to publication in Scopus

Robusteja tilastollisia menetelmiä hyvin moniulotteiselle datalle
Ollila, E. (Vastuullinen tutkija)
01/09/2016 → 31/12/2020
Projekti: Academy of Finland: Other research funding

Science-IT
Hakala, M. (Manager)
Perustieteiden korkeakoulu
Laitteistot/tilat: Facility

Siteeraa tätä

Mian, A., Raninen, E., & Ollila, E. (2020). A comparative study of supervised learning algorithms for symmetric positive definite features. teoksessa 28th European Signal Processing Conference, EUSIPCO 2020 - Proceedings (Sivut 950-954). Artikkeli 9287531 (European Signal Processing Conference). European Association For Signal and Image Processing. https://doi.org/10.23919/Eusipco47968.2020.9287531

@inproceedings{9dac28d0db07450596c2023f43f1ae60,

title = "A comparative study of supervised learning algorithms for symmetric positive definite features",

abstract = "In recent years, the use of Riemannian geometry has reportedly shown an increased performance for machine learning problems whose features lie in the symmetric positive definite (SPD) manifold. The present paper aims at reviewing several approaches based on this paradigm and provide a reproducible comparison of their output on a classic learning task of pedestrian detection. Notably, the robustness of these approaches to corrupted data will be assessed.",

keywords = "Covariance matrix, Pedestrian detection, Riemannian geometry, Supervised learning",

author = "Ammar Mian and Elias Raninen and Esa Ollila",

year = "2020",

doi = "10.23919/Eusipco47968.2020.9287531",

language = "English",

series = "European Signal Processing Conference",

publisher = "European Association For Signal and Image Processing",

pages = "950--954",

booktitle = "28th European Signal Processing Conference, EUSIPCO 2020 - Proceedings",

address = "Belgium",

note = "European Signal Processing Conference, EUSIPCO ; Conference date: 24-08-2020 Through 28-08-2020",

}

Mian, A, Raninen, E & Ollila, E 2020, A comparative study of supervised learning algorithms for symmetric positive definite features. julkaisussa 28th European Signal Processing Conference, EUSIPCO 2020 - Proceedings., 9287531, European Signal Processing Conference, European Association For Signal and Image Processing, Sivut 950-954, European Signal Processing Conference, Amsterdam, Alankomaat, 24/08/2020. https://doi.org/10.23919/Eusipco47968.2020.9287531

A comparative study of supervised learning algorithms for symmetric positive definite features. / Mian, Ammar; Raninen, Elias; Ollila, Esa.
28th European Signal Processing Conference, EUSIPCO 2020 - Proceedings. European Association For Signal and Image Processing, 2020. s. 950-954 9287531 (European Signal Processing Conference).

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

TY - GEN

T1 - A comparative study of supervised learning algorithms for symmetric positive definite features

AU - Mian, Ammar

AU - Raninen, Elias

AU - Ollila, Esa

N1 - Conference code: 28

PY - 2020

Y1 - 2020

N2 - In recent years, the use of Riemannian geometry has reportedly shown an increased performance for machine learning problems whose features lie in the symmetric positive definite (SPD) manifold. The present paper aims at reviewing several approaches based on this paradigm and provide a reproducible comparison of their output on a classic learning task of pedestrian detection. Notably, the robustness of these approaches to corrupted data will be assessed.

AB - In recent years, the use of Riemannian geometry has reportedly shown an increased performance for machine learning problems whose features lie in the symmetric positive definite (SPD) manifold. The present paper aims at reviewing several approaches based on this paradigm and provide a reproducible comparison of their output on a classic learning task of pedestrian detection. Notably, the robustness of these approaches to corrupted data will be assessed.

KW - Covariance matrix

KW - Pedestrian detection

KW - Riemannian geometry

KW - Supervised learning

UR - http://www.scopus.com/inward/record.url?scp=85099294575&partnerID=8YFLogxK

U2 - 10.23919/Eusipco47968.2020.9287531

DO - 10.23919/Eusipco47968.2020.9287531

M3 - Conference article in proceedings

AN - SCOPUS:85099294575

T3 - European Signal Processing Conference

SP - 950

EP - 954

BT - 28th European Signal Processing Conference, EUSIPCO 2020 - Proceedings

PB - European Association For Signal and Image Processing

T2 - European Signal Processing Conference

Y2 - 24 August 2020 through 28 August 2020

ER -

Mian A, Raninen E, Ollila E. A comparative study of supervised learning algorithms for symmetric positive definite features. julkaisussa 28th European Signal Processing Conference, EUSIPCO 2020 - Proceedings. European Association For Signal and Image Processing. 2020. s. 950-954. 9287531. (European Signal Processing Conference). doi: 10.23919/Eusipco47968.2020.9287531

A comparative study of supervised learning algorithms for symmetric positive definite features

Abstrakti

Julkaisusarja

Conference

Pääsy asiakirjaan

Muut tiedostot ja linkit

Sormenjälki

Projektit

Robusteja tilastollisia menetelmiä hyvin moniulotteiselle datalle

Laitteet

Science-IT

Siteeraa tätä