|
Work item:
|
F.MACVS
|
|
Subject/title:
|
Requirements and framework of multimodal agents collaboration based video surveillance system
|
|
Status:
|
Under study
|
|
Approval process:
|
AAP
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
New
|
|
Equivalent number:
|
-
|
|
Timing:
|
2027 (Medium priority)
|
|
Liaison:
|
-
|
|
Supporting members:
|
State Grid Corporation of China, China Telecom, Zhejiang Dahua Technology Co., Ltd., Zhejiang Lab
|
|
Summary:
|
The increasing demand for intelligent, adaptive, and context-aware video surveillance systems have outpaced the capabilities of traditional intelligent visual surveillance frameworks. Conventional systems—typically relying on static visual analysis and centralized processing—are no longer sufficient to meet the challenges of modern surveillance environments, which require real-time responsiveness, cross-domain adaptability, and multi-source data integration. Recent advancements in AI agent technologies and multi-modal fusion have enabled the development of distributed and collaborative surveillance systems. These systems leverage autonomous AI agents to perform specialized tasks such as video analysis, behaviour recognition, event correlation, and situation reasoning. By incorporating data beyond video—such as audio streams, textual metadata, and environmental sensor inputs—these systems achieve more robust and accurate situational awareness.
This Recommendation specifies the requirements and framework for multimodal agents collaboration based video surveillance systems, focusing on agent roles, functional components, multi-agent collaboration, and multi-modal integration. It aims to promote scalability, modularity, and interoperability to support trustworthy, real-time, and intelligent surveillance across a wide range of application domains.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2025-11-17 16:24:31
|
|
Last update:
2025-11-18 17:44:44
|
|
