VR全景视频渲染输出安全性自动评价模型

Automatic Evaluation Model for Output Security of VR Panoramic Video Rendering

解决当前模型相对误差较大以及评价耗时较长的问题,设计并提出一种VR全景视频渲染输出安全性自动评价模型。结合感知重要性度量提取VR全景视频特征;通过重要性排列结果获取VR全景视频的感知信息以及重要性度量;计算失真视频与参考感知视频的差异,获取VR全景视频渲染输出过程中的失真度量以及质量值,结合计算结果完成VR全景视频渲染输出安全性自动评价模型的构建。仿真实验结果表明,所提模型能够有效降低相对误差,评价耗时较低。通过随机神经网络训练模型,以达到最终视频输出安全性自动评价的目的。

In order to solve the problems of large relative error and long evaluation time, an automatic evaluation model of VR panoramic video rendering output security is designed and proposed. It extracts the features of VR panoramic video by combining the perceptual importance measure;obtains the perceptual information and importance measure of VR panoramic video through the importance ranking results;calculates the difference between distorted video and reference perceived video to obtain the distortion measurement and quality value in the process of VR panoramic video rendering and output, and combines the calculation results to complete the VR panoramic video rendering output security automatic evaluation model Build. The model is trained by stochastic neural network to achieve the purpose of automatic evaluation of final video output security. Simulation results show that the proposed model can effectively reduce the relative error and the evaluation time is low.