Academy of Interdisciplinary Studies Division of Emerging Interdisciplinary Areas 201 Machine Learning on Wearable Devices Supervisor: HUI, Pan / EMIA Student: GUO, Zhimao / COMP Course: UROP1100, Fall This semester I have been working in the lab MetaHKUST under the supervision of professor Pan Hui. MetaHKUST is a project of building an immersive metaverse of HKUST campus. Metaverse is a really hot concept in recent years, and it involves many technologies in many fields. I have been working on the interactions of avatars in Virtual Environments. Normally users are entering metaverse with a Virtual Reality (VR) head-mounted displays (HMD) and 2 controllers. The number of sensors is very limited, and it can only detect the movements and positions of head and hands. I was working on the project to find a solution under this situation. I am responsible to collect related data and test open-sourced codes. This report I will illustrate what I did and what I learnt. Machine Learning on Wearable Devices Supervisor: HUI, Pan / EMIA Student: WU, Ze / COMP Course: UROP1100, Fall With the rapid development of Virtual Reality(VR) related area, high GPU performance becomes an important requirement of VR devices. With the limited hardware in VR devices, Cloud gaming becomes a suitable solution for these requirements. However, reliable low latency video transmission is difficult to achieve in normal network environment because of unstable network state. In most of SEVER-CLIENT video transmission model, fixed encode bit rate are used. However, fixed encode bit rate leads to a high game latency when the network latency is high. That comes out the adaptive bit rate video transmission model in VR cloud gaming, which minimizes the game latency and gives a high quality video transmission.
RkJQdWJsaXNoZXIy NDk5Njg=