|
Work item:
|
F.RFVFMP
|
|
Subject/title:
|
Requirements and framework for the vision foundation models platform
|
|
Status:
|
Under study [Issued from previous study period]
|
|
Approval process:
|
AAP
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
New
|
|
Equivalent number:
|
-
|
|
Timing:
|
2026-Q4 (Medium priority)
|
|
Liaison:
|
ISO/IEC JTC1 SC42
|
|
Supporting members:
|
China Telecom, China Unicom, Zhejiang Dahua Technology, MIIT
|
|
Summary:
|
The vision foundation models enable the machines to enhance the perception and understanding of the visual information, such as enabling generalization, robustness and higher-order reasoning. The development of vision foundation models enhances the capabilities of information and communication technology, giving rise to numerous innovative applications and services. These models have tremendous potential and value in fields such as autonomous driving, manufacturing industry, and robotics. The vision foundation models platform is a software platform used for developing, training, deploying, and enabling the utilization of vision foundation models, with functionalities such as data management, model training, model delivery, and model selection.
This Recommendation identifies requirements and framework for the vision foundation models platform. It provides the introduction of vision foundation models platform, identifies the requirements and framework for a vision foundation models platform, and specifies the corresponding workflow. Use cases are also provided in an appendix. The specification of vision foundation model is out of the scope of this Recommendation. This Recommendation is intended to guide the construction and operation of vision foundation models platform.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2024-02-02 15:50:40
|
|
Last update:
2025-03-05 13:35:00
|
