Effective and Efficient Similarity Searching in Motion Capture
Data

J 2018

Effective and Efficient Similarity Searching in Motion Capture Data

SEDMIDUBSKÝ, Jan, Petr ELIÁŠ a Pavel ZEZULA

Základní údaje

Originální název

Effective and Efficient Similarity Searching in Motion Capture Data

Autoři

SEDMIDUBSKÝ, Jan (203 Česká republika, garant, domácí), Petr ELIÁŠ (203 Česká republika, domácí) a Pavel ZEZULA (203 Česká republika, domácí)

Vydání

Multimedia Tools and Applications, Springer US, 2018, 1380-7501

Další údaje

Jazyk

angličtina

Typ výsledku

Článek v odborném periodiku

Obor

10201 Computer sciences, information science, bioinformatics

Stát vydavatele

Spojené státy

Utajení

není předmětem státního či obchodního tajemství

Impakt faktor

Impact factor: 2.101

Kód RIV

RIV/00216224:14330/18:00100703

Organizační jednotka

Fakulta informatiky

DOI

http://dx.doi.org/10.1007/s11042-017-4859-7

UT WoS

000433202100021

Klíčová slova anglicky

Motion capture data retrieval;Effective similarity measure;Efficient indexing;k-NN query;Motion image;Convolutional neural network;Fixed-size motion feature

Štítky

DISA

Příznaky

Mezinárodní význam, Recenzováno

Změněno: 16. 4. 2019 07:37, doc. RNDr. Jan Sedmidubský, Ph.D.

Anotace

V originále

Motion capture data describe human movements in the form of spatio-temporal trajectories of skeleton joints. Intelligent management of such complex data is a challenging task for computers which requires an effective concept of motion similarity. However, evaluating the pair-wise similarity is a difficult problem as a single action can be performed by various actors in different ways, speeds or starting positions. Recent methods usually model the motion similarity by comparing customized features using distance-based functions or specialized machine-learning classifiers. By combining both these approaches, we transform the problem of comparing motions of variable sizes into the problem of comparing fixed-size vectors. Specifically, each rather-short motion is encoded into a compact visual representation from which a highly descriptive 4,096-dimensional feature vector is extracted using a fine-tuned deep convolutional neural network. The advantage is that the fixed-size features are compared by the Euclidean distance which enables efficient motion indexing by any metric-based index structure. Another advantage of the proposed approach is its tolerance towards an imprecise action segmentation, the variance in movement speed, and a lower data quality. All these properties together bring new possibilities for effective and efficient large-scale retrieval.

Návaznosti

GBP103/12/G084, projekt VaV

Název: Centrum pro multi-modální interpretaci dat velkého rozsahu

Investor: Grantová agentura ČR, Centrum pro multi-modální interpretaci dat velkého rozsahu

MUNI/A/0992/2016, interní kód MU

Název: Zapojení studentů Fakulty informatiky do mezinárodní vědecké komunity (Akronym: SKOMU)

Investor: Masarykova univerzita, Zapojení studentů Fakulty informatiky do mezinárodní vědecké komunity, DO R. 2020_Kategorie A - Specifický výzkum - Studentské výzkumné projekty

Citovat

SEDMIDUBSKÝ, Jan, Petr ELIÁŠ a Pavel ZEZULA. Effective and Efficient Similarity Searching in Motion Capture Data. Multimedia Tools and Applications. Springer US, 2018, roč. 77, č. 10, s. 12073-12094. ISSN 1380-7501. Dostupné z: https://dx.doi.org/10.1007/s11042-017-4859-7.

@article{1381296,
   author = {Sedmidubský, Jan and Eliáš, Petr and Zezula, Pavel},
   article_number = {10},
   doi = {http://dx.doi.org/10.1007/s11042-017-4859-7},
   keywords = {Motion capture data retrieval;Effective similarity measure;Efficient indexing;k-NN query;Motion image;Convolutional neural network;Fixed-size motion feature},
   language = {eng},
   issn = {1380-7501},
   journal = {Multimedia Tools and Applications},
   title = {Effective and Efficient Similarity Searching in Motion Capture Data},
   volume = {77},
   year = {2018}
}

TY  - JOUR
ID  - 1381296
AU  - Sedmidubský, Jan - Eliáš, Petr - Zezula, Pavel
PY  - 2018
TI  - Effective and Efficient Similarity Searching in Motion Capture Data
JF  - Multimedia Tools and Applications
VL  - 77
IS  - 10
SP  - 12073-12094
EP  - 12073-12094
PB  - Springer US
SN  - 13807501
KW  - Motion capture data retrieval;Effective similarity measure;Efficient indexing;k-NN query;Motion image;Convolutional neural network;Fixed-size motion feature
N2  - Motion capture data describe human movements in the form of spatio-temporal trajectories of skeleton joints. Intelligent management of such complex data is a challenging task for computers which requires an effective concept of motion similarity. However, evaluating the pair-wise similarity is a difficult problem as a single action can be performed by various actors in different ways, speeds or starting positions. Recent methods usually model the motion similarity by comparing customized features using distance-based functions or specialized machine-learning classifiers. By combining both these approaches, we transform the problem of comparing motions of variable sizes into the problem of comparing fixed-size vectors. Specifically, each rather-short motion is encoded into a compact visual representation from which a highly descriptive 4,096-dimensional feature vector is extracted using a fine-tuned deep convolutional neural network. The advantage is that the fixed-size features are compared by the Euclidean distance which enables efficient motion indexing by any metric-based index structure. Another advantage of the proposed approach is its tolerance towards an imprecise action segmentation, the variance in movement speed, and a lower data quality. All these properties together bring new possibilities for effective and efficient large-scale retrieval.
ER  -

SEDMIDUBSKÝ, Jan, Petr ELIÁŠ a Pavel ZEZULA. Effective and Efficient Similarity Searching in Motion Capture Data. \textit{Multimedia Tools and Applications}. Springer US, 2018, roč.~77, č.~10, s.~12073-12094. ISSN~1380-7501. Dostupné z: https://dx.doi.org/10.1007/s11042-017-4859-7.

Podrobný výpis o publikaci