|
Work item:
|
P.obj-recog
|
|
Subject/title:
|
Object-recognition-rate-estimation model in surveillance video of autonomous driving
|
|
Status:
|
Under study
|
|
Approval process:
|
AAP
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
Rev.
|
|
Equivalent number:
|
-
|
|
Timing:
|
2025 (High priority)
|
|
Liaison:
|
-
|
|
Supporting members:
|
NTT, Telefon AB - LM Ericsson
|
|
Summary:
|
To ensure the safety of autonomous driving, objects that interfere with driving need to be automatically recognized. To do that, an object-recognition system is needed to support autonomous driving, such as autonomous braking and passing. In general, the object recognition is performed automatically in autonomous driving systems using cameras mounted on an autonomous car. Under certain conditions such as on a highway, autonomous driving systems work well without any human support. On the other hand, current autonomous driving systems are difficult to use on local streets because there are various objects or people such as traffic signs or pedestrians. To use the current autonomous driving systems on local streets, a remote monitoring system has also been studied in which an observer can recognize objects and brake remotely.
In the remote monitoring system, surveillance video is encoded and transmitted via radio access networks. Therefore, the encoding video bitrate is varied, and the packet loss occurs due to fluctuations in the network bandwidth. These factors affect the object-recognition rate. Characteristics of objects and other factors in surveillance video (e.g., color, size, and background of objects) also affect the object-recognition rate. Therefore, a new technique needs to be established to estimate the object-recognition rate of surveillance video.
In this work item, an object-recognition-rate-estimation model will be developed for the surveillance video of autonomous driving. To determine object-recognition rate, the subjective assessment methods for recognition task described in Recommendation P.912 can be applied.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2023-02-06 10:30:52
|
|
Last update:
2024-04-25 15:58:01
|
