Using virtual reality technology to construct computer-aided animation material development

Computer-assisted virtual reality animation has created a new form of animation with its subversive visual experience. This paper proposes a dynamic mapping and load balancing strategy between multi-granularity tasks and heterogeneous parallel devices. A parallel task based on OpenCL for heterogeneous systems of CPU and GPU-Disturbing Particle Animation System is designed. We port existing CUDA-based parallel algorithms to OpenCL, thus breaking through the hardware limitations, and improving the portability and versatility of parallel algorithms. Combining the particle system animation material and the skeletal animation above, we design the principle of multi-tasking and heterogeneous parallel device mapping. According to the different OpenCL parallel acceleration capabilities of the CPU and GPU for different tasks, parallel tasks are allocated according to the acceleration ratio coefficient, and the performance is improved by reducing the waiting time of the device. This paper determines the relative optimal parameters of the random forest model through experiments, and through a comparison experiment with the decision tree model, shows the performance advantages of the random forest model that we designed and implemented for the automatic generation of animation. We counted and tested 614 animation materials, and finally rendered 379 animation materials. Among them, the animation active learning system was called when processing 369 animation materials, which showed the feasibility of the animation active learning system. The probability that the user likes the scene to be selected is improved, and the probability that the user does not like the scene to be selected is reduced. This shows the effectiveness of the computer-aided virtual reality animation system. © 2022 CAD Solutions, LLC.