研究目的
To increase the sense of immersion of 360 virtual reality (VR) images by proposing and implementing foveated rendering technology through precise ROI detection using eye-tracking-based head-mounted display.
研究成果
The implemented foveated rendering technology shows high rendering speeds and high-quality textures, improving the performance of the texture rendering framerate by 10–15% on average compared with the case of applying only the existing tiled streaming technology, and up to 18% depending on the eye-tracking coordinate detection and timing.
研究不足
The study does not mention the specific models of the HMD devices used or the exact parameters of the eye-tracking technology. The improvement in rendering performance is dependent on the eye-tracking coordinate detection and timing.
1：Experimental Design and Method Selection:
The study uses HEVC tiled video-based image-decoding and -rendering method for foveated rendering.
2：Sample Selection and Data Sources:
Eye tracking coordinates (x, y, z) in 3D space were extracted using an HMD device that supports eye tracking.
3：List of Experimental Equipment and Materials:
HMD device supporting eye tracking, MPEG-DASH standard player.
4：Experimental Procedures and Operational Workflow:
The player receives eye-tracking coordinates through the HMD, detects the ROI, and defines the FOV based on the ROI. The adaptation set is constructed with two high-definition and low-definition tiles for the FOV and non-FOV areas, respectively.
5：Data Analysis Methods:
The performance of the texture rendering framerate in frames per second (FPS) was compared.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
