Jiangkai Wu (吴 将凯)
Affiliations. Peking University, Beijing, China.

I am a Ph.D. student at Peking University majoring in Computer Science, advised by Prof. Xinggong Zhang. My current research interest is applying generative vision and graphics to video streaming to optimize system performance. My work involves algorithm design and implementation optimization, spanning computer networks, vision, graphics, and machine learning.
My current research interest:
- Low bandwidth: Employing generative large models (such as Sora, Stable Diffusion) to inverse-generate low bitrate representations for videos.
- Low latency: Using predicted frames to instantly respond to user actions, achieving a remote interaction system with low motion-to-photon latency.
- High immersion: Building a high-freedom, high-fidelity video streaming system based on NeRF or 3DGS, optimizing perceptual quality, efficiency and bitrate.
Contact me at: jiangkai.wu@stu.pku.edu.cn
selected publications
- APNetPromptMobile: Efficient Promptus for Low Bandwidth Mobile Video StreamingAccepted by APNet, 2025
- ICME3DGCoding: Novel Framework for 3D Gaussian Video Incremental Training and CodingAccepted by ICME, 2025
- Under ReviewGecko: High-Quality Video Streaming via Generative Prompt ChunksUnder Review, 2025
- ICCSRFC: Scalable Radiance Fields Streaming with Planar CodecIn ICC 2024-IEEE International Conference on Communications, 2024
- EMSLow-bitrate Volumetric Video Streaming with Depth ImageIn Proceedings of the 2024 SIGCOMM Workshop on Emerging Multimedia Systems, 2024