Case Studies
Virtual avatars for AMD GPUs in video games streaming
About the service
A hardware manufacturer required an AMD-optimized application that utilized dedicated GPUs for game rendering and the integrated GPU for executing AI tasks. The implemented solution employed Edge AI and GStreamer to perform real-time facial gesture extraction, background removal, and 3D avatar rendering from live webcam feeds. This architecture delivered expressive, low-latency overlays without impacting gameplay performance, ensuring smooth operation even during high-resolution, resource-intensive gaming sessions.
What We Achieved
The solution consisted of a visual webcam system that modified and merged two independent video streams. It employed dual GPU processing for AI inference and game execution, including hardware-accelerated and highly accurate facial feature detection with 3D rendering of visual avatars.
Real-time background removal and 3D avatar rendering enhanced the visual experience by eliminating distractions and providing expressive, animated characters without needing green screens. The system streamed triple-A games with the modified video feed while maintaining smooth performance through an efficient dual GPU architecture optimized for AMD hardware. This approach delivered low-latency, high-quality visuals with minimal GPU usage, ensuring consistent high-resolution content delivery.
