Interactive Retrieval System for Multi-Stream Collections: multiXview at CASTLE 2025 Interactive Grand Challenge

Omar Shahbaz Khan; Ujjwal Sharma; Gonçalo Marcelino; Aaron Duane; Stevan Rudinac; Marcel Worring; Björn Þór Jónsson

doi:10.1145/3746027.3760243

Interactive Retrieval System for Multi-Stream Collections: multiXview at CASTLE 2025 Interactive Grand Challenge

Omar Shahbaz Khan, Ujjwal Sharma, Gonçalo Marcelino, Aaron Duane, Stevan Rudinac, Marcel Worring, Björn Þór Jónsson

Publikation: Artikel i tidsskrift og konference artikel i tidsskrift › Konferenceartikel › Forskning › peer review

Abstract

We introduce multiXview, an interactive retrieval framework for synchronized multi-camera video collections. It features a multi-index search engine that supports natural-language queries over visual embeddings, speech transcripts, and scene descriptions. It supports a synchronized multi-stream player offering parallel playback, and a timeline-based navigation view for temporal scoping and faceted exploration. These components address the redundancy and fragmentation of overlapping egocentric and exocentric video feeds and enable users to locate, aggregate, and reconstruct events across partial perspectives. This paper focuses on system design and implementation, with quantitative and qualitative evaluation to take place at the CASTLE 2025 Grand Challenge Interactive Track.

Originalsprog	Engelsk
Tidsskrift	MM '25
Sider (fra-til)	14273–14279
Antal sider	7
DOI	https://doi.org/10.1145/3746027.3760243
Status	Udgivet - okt. 2025
Udgivet eksternt	Ja
Begivenhed	Multimedia - Dublin, Irland Varighed: 27 okt. 2025 → 31 okt. 2025 Konferencens nummer: 33 https://acmmm2025.org/

Konference

Konference	Multimedia
Nummer	33
Land/Område	Irland
By	Dublin
Periode	27/10/2025 → 31/10/2025
Internetadresse	https://acmmm2025.org/

Emneord

castle
multi-stream interface
multimedia analytics
multimedia retrieval

Adgang til dokumentet

10.1145/3746027.3760243Licens: CC BY-SA

Citationsformater

@inproceedings{b94b9508dcb643e0b289b15dbf92e28f,

title = "Interactive Retrieval System for Multi-Stream Collections: multiXview at CASTLE 2025 Interactive Grand Challenge",

abstract = "We introduce multiXview, an interactive retrieval framework for synchronized multi-camera video collections. It features a multi-index search engine that supports natural-language queries over visual embeddings, speech transcripts, and scene descriptions. It supports a synchronized multi-stream player offering parallel playback, and a timeline-based navigation view for temporal scoping and faceted exploration. These components address the redundancy and fragmentation of overlapping egocentric and exocentric video feeds and enable users to locate, aggregate, and reconstruct events across partial perspectives. This paper focuses on system design and implementation, with quantitative and qualitative evaluation to take place at the CASTLE 2025 Grand Challenge Interactive Track.",

keywords = "castle, multi-stream interface, multimedia analytics, multimedia retrieval, Multimedia Retrieval, Multimedia Analytics, Multi-Stream Interface, CASTLE",

author = "Khan, \{Omar Shahbaz\} and Ujjwal Sharma and Gon{\c c}alo Marcelino and Aaron Duane and Stevan Rudinac and Marcel Worring and J{\'o}nsson, \{Bj{\"o}rn {\TH}{\'o}r\}",

year = "2025",

month = oct,

doi = "10.1145/3746027.3760243",

language = "English",

pages = "14273–14279",

journal = "MM '25",

note = "Multimedia , MM 25 ; Conference date: 27-10-2025 Through 31-10-2025",

url = "https://acmmm2025.org/",

}

TY - GEN

T1 - Interactive Retrieval System for Multi-Stream Collections: multiXview at CASTLE 2025 Interactive Grand Challenge

AU - Khan, Omar Shahbaz

AU - Sharma, Ujjwal

AU - Marcelino, Gonçalo

AU - Duane, Aaron

AU - Rudinac, Stevan

AU - Worring, Marcel

AU - Jónsson, Björn Þór

N1 - Conference code: 33

PY - 2025/10

Y1 - 2025/10

N2 - We introduce multiXview, an interactive retrieval framework for synchronized multi-camera video collections. It features a multi-index search engine that supports natural-language queries over visual embeddings, speech transcripts, and scene descriptions. It supports a synchronized multi-stream player offering parallel playback, and a timeline-based navigation view for temporal scoping and faceted exploration. These components address the redundancy and fragmentation of overlapping egocentric and exocentric video feeds and enable users to locate, aggregate, and reconstruct events across partial perspectives. This paper focuses on system design and implementation, with quantitative and qualitative evaluation to take place at the CASTLE 2025 Grand Challenge Interactive Track.

AB - We introduce multiXview, an interactive retrieval framework for synchronized multi-camera video collections. It features a multi-index search engine that supports natural-language queries over visual embeddings, speech transcripts, and scene descriptions. It supports a synchronized multi-stream player offering parallel playback, and a timeline-based navigation view for temporal scoping and faceted exploration. These components address the redundancy and fragmentation of overlapping egocentric and exocentric video feeds and enable users to locate, aggregate, and reconstruct events across partial perspectives. This paper focuses on system design and implementation, with quantitative and qualitative evaluation to take place at the CASTLE 2025 Grand Challenge Interactive Track.

KW - castle

KW - multi-stream interface

KW - multimedia analytics

KW - multimedia retrieval

KW - Multimedia Retrieval

KW - Multimedia Analytics

KW - Multi-Stream Interface

KW - CASTLE

U2 - 10.1145/3746027.3760243

DO - 10.1145/3746027.3760243

M3 - Conference article

SP - 14273

EP - 14279

JO - MM '25

JF - MM '25

T2 - Multimedia

Y2 - 27 October 2025 through 31 October 2025

ER -

Interactive Retrieval System for Multi-Stream Collections: multiXview at CASTLE 2025 Interactive Grand Challenge

Abstract

Konference

Emneord

Adgang til dokumentet

Fingeraftryk

Citationsformater