Realization and Experimental Research on Immersive Visual Communication System Based on Real-Time Rendering

Aiming at the problems of low image processing efficiency and insufficient rendering quality in virtual reality content creation, a real-time rendering system based on deep learning is designed and realized. An improved U-Net network structure is used for image processing, and a rendering speed of 112fps is realized in a standard test environment, with the end-to-end delay controlled within 12ms. The system achieves a task completion rate of 95.8% and a satisfaction score of 4.6 in a test with 1000 users. The experimental results show that the system significantly improves the rendering efficiency and interactive experience, and provides reliable technical support for virtual reality content creation.