|
Work item:
|
F.AI-RSRSreqs
|
|
Subject/title:
|
Requirements for real-time super-resolution service based on artificial intelligence
|
|
Status:
|
Under study [Issued from previous study period]
|
|
Approval process:
|
-
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
New
|
|
Equivalent number:
|
-
|
|
Timing:
|
2026 (Medium priority)
|
|
Liaison:
|
ITU-T SG9, IEEE
|
|
Supporting members:
|
China Telecom, China Mobile, ZTE, MIIT/CAICT
|
|
Summary:
|
With the explosive growth development of video services, users' requirements for the resolution and quality of real-time image/video services is gradually improving. Taking the monitoring service as an example, the quality of service is reduced due to the insufficient acquisition capability of video monitoring terminal, image quality damage caused by network transmission and low resolution. For low resolution (LR) videos, the artificial intelligence based super-resolution (SR) technology can ensure the detail texture of the video and improve the video resolution to the greatest extent. The current SR technology is still limited to offline video repairing. Using the conventional neural network, one of the artificial intelligence algorithms, the LR images/videos can be reconstructed and the image and video quality improve practically.
This document provides the requirements for real-time super-resolution service based on artificial intelligence.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2021-05-21 17:41:24
|
|
Last update:
2025-03-18 17:01:09
|
