2022 Fiscal Year Research-status Report
Towards quality monitoring and managing for online learning
| Project/Area Number |
22K12299
|
|---|
| Research Institution | The University of Aizu |
|---|
Principal Investigator |
Truong CongThang 会津大学, コンピュータ理工学部, 上級准教授 (40622957)
|
|---|
| Project Period (FY) |
2022-04-01 – 2025-03-31
|
| Keywords | Quality of experience / Adaptive streaming / Online learning / Quality model / Media analysis / Multi-feature learning |
|---|
| Outline of Annual Research Achievements |
In AY2022, we investigated problems of Quality of Experience (QoE) and video delivery for online learning. We evaluated of a large number of quality models for both PC and mobile users. Because online learning is mostly received via PC displays and has long durations, we found that just a few of quality models were appropriate for measuring the QoE at the learner’s side. Besides, the potential of multi-feature learning approach was studied for modeling. For video delivery, a new solution was proposed to send the same video to multiple users. Our solution employed scalable video coding and multicast mechanism to effectively deliver a video to many users in a session. We also investigated key issues and techniques in system implementation for QoE management for post-pandemic online learning.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
In the last year, we have successfully finished the investigation of both quality models and media delivery mechanism. The research has progressed rather smoothly as planned. It is found that existing quality models are still limited and need significant improvements. Thus, our current focus is on developing effective quality models to support both good perception and good semantic understanding for users.
|
| Strategy for Future Research Activity |
In the AY2023, we will focus on the development of new quality models. The main approach is to take into account multiple features of media contents. Because in online learning, the quality of media contents may affect both the perception and understanding of users, the features will be extracted by considering both the perceptual aspect and semantic aspect of contents. New deep learning methods will be applied for efficient and effective deployment of quality models.
|
| Causes of Carryover |
Amount to be used next Fiscal Year（B-A）is just 146 yen. This amount will be used together with the allocated amount in FY2023 to buy articles for the research.
|
Research Products
(5 results)
