Filtering Few-Level Segment Regions for Efficient Subsequence
Search in 3D Human Motions

D 2026

Filtering Few-Level Segment Regions for Efficient Subsequence Search in 3D Human Motions

ČERNEK, Andrej a Jan SEDMIDUBSKÝ

Základní údaje

Originální název

Filtering Few-Level Segment Regions for Efficient Subsequence Search in 3D Human Motions

Autoři

ČERNEK, Andrej a Jan SEDMIDUBSKÝ

Vydání

48th European Conference on Information Retrieval (ECIR), 16 s. 2026

Nakladatel

Springer

Další údaje

Jazyk

angličtina

Typ výsledku

Stať ve sborníku

Obor

10200 1.2 Computer and information sciences

Forma vydání

elektronická verze "online"

Organizační jednotka

Fakulta informatiky

Klíčová slova anglicky

3D skeleton sequences; subsequence search; overlapping segmentation; query localization; 3D action retrieval; motion similarity

Štítky

DISA

Příznaky

Mezinárodní význam, Recenzováno

Změněno: 17. 12. 2025 18:16, doc. RNDr. Jan Sedmidubský, Ph.D.

Anotace

V originále

Efficient localization of query-similar subsequences in a database of untrimmed 3D human motion data is crucial to applications in numerous domains. We propose a novel subsequence search approach that partitions untrimmed database motions into segments across a few levels to accommodate variably-sized queries, addressing the limitations of single- and many-level state-of-the-art methods. By determining a deep similarity between the query and database segments, we specifically identify larger regions within the database motions likely to contain query-similar subsequences. These regions are then narrowly examined to determine the precise location of relevant subsequences, considering also variations in execution speed. While this approach contributes to a high retrieval quality, it also requires high search costs. Therefore, we propose two filtering techniques that further decrease the number of examined subsequences by more than an order of magnitude on a newly established benchmark across four challenging PKU-MMD sub-datasets.

Návaznosti

GF23-07040K, projekt VaV

Název: Naučené indexy pro podobností hledání

Investor: Grantová agentura ČR, Naučené indexy pro podobností hledání, Lead agentura

Citovat

ČERNEK, Andrej a Jan SEDMIDUBSKÝ. Filtering Few-Level Segment Regions for Efficient Subsequence Search in 3D Human Motions. Online. In 48th European Conference on Information Retrieval (ECIR). Springer, 2026, 16 s.

@inproceedings{2541320,
   author = {Černek, Andrej and Sedmidubský, Jan},
   booktitle = {48th European Conference on Information Retrieval (ECIR)},
   keywords = {3D skeleton sequences; subsequence search; overlapping segmentation; query localization; 3D action retrieval; motion similarity},
   howpublished = {elektronická verze "online"},
   language = {eng},
   publisher = {Springer},
   title = {Filtering Few-Level Segment Regions for Efficient Subsequence Search in 3D Human Motions},
   year = {2026}
}

TY  - CONF
ID  - 2541320
AU  - Černek, Andrej - Sedmidubský, Jan
PY  - 2026
TI  - Filtering Few-Level Segment Regions for Efficient Subsequence Search in 3D Human Motions
PB  - Springer
KW  - 3D skeleton sequences
KW  - subsequence search
KW  - overlapping segmentation
KW  - query localization
KW  - 3D action retrieval
KW  - motion similarity
N2  - Efficient localization of query-similar subsequences in a database of untrimmed 3D human motion data is crucial to applications in numerous domains. We propose a novel subsequence search approach that partitions untrimmed database motions into segments across a few levels to accommodate variably-sized queries, addressing the limitations of single- and many-level state-of-the-art methods. By determining a deep similarity between the query and database segments, we specifically identify larger regions within the database motions likely to contain query-similar subsequences. These regions are then narrowly examined to determine the precise location of relevant subsequences, considering also variations in execution speed. While this approach contributes to a high retrieval quality, it also requires high search costs. Therefore, we propose two filtering techniques that further decrease the number of examined subsequences by more than an order of magnitude on a newly established benchmark across four challenging PKU-MMD sub-datasets.
ER  -

ČERNEK, Andrej a Jan SEDMIDUBSKÝ. Filtering Few-Level Segment Regions for Efficient Subsequence Search in 3D Human Motions. Online. In \textit{48th European Conference on Information Retrieval (ECIR)}. Springer, 2026, 16 s.

Přehled o publikaci