2022 Fiscal Year Annual Research Report
AI-empowered Point Cloud Video Streaming
| Project/Area Number |
20H04174
|
|---|
| Research Institution | The University of Electro-Communications |
|---|
Principal Investigator |
劉 志 電気通信大学, 大学院情報理工学研究科, 准教授 (90750240)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
太田 香 室蘭工業大学, 大学院工学研究科, 文部科学省卓越研究員(教授) (50713971)
李 鶴 室蘭工業大学, 大学院工学研究科, 文部科学省卓越研究員(准教授) (40759891)
Kien Nguyen 千葉大学, 大学院工学研究院, 准教授 (80647222)
|
|---|
| Project Period (FY) |
2020-04-01 – 2024-03-31
|
| Keywords | point cloud / video streaming / MEC / AI / VR/AR/MR |
|---|
| Outline of Annual Research Achievements |
Point cloud video is the most popular representation of hologram and volumetric video, which is the medium to precedent natural content in VR/AR/MR, and is expected to be the next generation video by providing users with better immersive viewing experience. Point cloud video has wide applications in Society 5.0 such as for online education and entertainment. To further enhance these applications, networked point cloud video is in critical demand. This project aims to solve the core technical questions towards a high quality point cloud video streaming and promote the applications of point cloud video in Society 5.0.
In the past year, we have developed AI-based streaming algorithms that deliver enhanced video quality to users. Additionally, we have optimized the MEC to provide short-delay additional computation capability. To evaluate performance, we have created prototypes. Some of our findings have been published or submitted to IEEE/ACM journals and conferences.
|
| Current Status of Research Progress |
Current Status of Research Progress
1: Research has progressed more than it was originally planned.
Reason
非常に良い結果を得ることができ、重要な問題を解決することができました。
|
| Strategy for Future Research Activity |
Next year I will focus on the following two technical challenges:
1. AI-empowered point cloud video transmission: We will use AI technologies (such as transformer, LSTM,) to predict the user viewing angle and efficiently and adaptively utilize nearby available communication and computing resources (including the resource provided MEC) to provide required computation for encoding, transcoding and decoding, where existing resource allocation schemes cannot be directly applied.
2. Prototype building and scheme adjustment: The proposed point cloud video streaming system will be implemented based on off-the-shelf devices and the proposed schemes will be adjusted according to their performance. We will also summarize the results and conclude this project.
|
