۱- Introduction

۲- The framework for omnidirectional video coding

۳- Omnidirectional video projections

۴- Omnidirectional video quality evaluation

۵- The comparison of projections and evaluation criteria

۶- Discussions and conclusions

References

Abstract

In this paper, we review the recent advances in the pipeline of omnidirectional video processing including projection and evaluation. Being distinct from the traditional video, the omnidirectional video, also called panoramic video or 360 degree video, is in the spherical domain, thus specialized tools are necessary. For this type of video, each picture should be projected to a 2-D plane for encoding and decoding, adapting to the input of existing video coding systems. Thus the coding influence of the projection and the accuracy of the evaluation method are very important in this pipeline. Recent advances, such as different projection methods benefiting video coding, specialized video quality evaluation metrics and optimized methods for transmission, are all presented and classified in this paper. In addition, the coding performances under different projection methods are specified. The future trends of omnidirectional video processing are also discussed.

Introduction

Pursuing the immersive experience to simulate the real world in the digital devices has been an increasingly hot topic. Many efforts are in the way to provide better user experience with high resolution/quality video, HDR video content, large screen display, etc. Recently, with the availability of commercial Virtual Reality (VR) Head Mounted Displays (HMD) such as Oculus Rift or HTC Vive, VR video application attracts great attention. With these products, users can enjoy the omnidirectional video and can choose their desired viewport by moving heads as they do in the real world, thus the immersive experience can be provided. As the content of VR, the demand of omnidirectional video proliferates with the increasing attraction and popularity of VR applications, while it should be noted that there still exist many obstacles for omnidirectional video processing. For immersive visual experience, high resolution (6K or beyond) and high frame rate (e.g., 90 fps) are expected, so that the bitstream tends to be very large, causing severe resource consuming on storage and bandwidth. Therefore, improving compression efficiency of omnidirectional video is in urgent demand.