Energy-Efficient Video Processing for Virtual Reality

Abstract

Virtual reality (VR) has huge potential to enable radically new applications, behind which spherical panoramic video processing is one of the backbone techniques. However, current VR systems reuse the techniques designed for processing conventional planar videos, resulting in significant energy inefficiencies. Our characterizations show that operations that are unique to processing 360° VR content constitute 40% of the total processing energy consumption. We present EVR, an end-to-end system for energy-efficient VR video processing. EVR recognizes that the major contributor to the VR tax is the projective transformation (PT) operations. EVR mitigates the overhead of PT through two key techniques: semantic-aware streaming (SAS) on the server and hardware-accelerated rendering (HAR) on the client device. EVR uses SAS to reduce the chances of executing projective transformation on VR devices by pre-rendering 360° frames in the cloud. Different from conventional pre-rendering techniques, SAS exploits the key semantic information inherent in VR content that is previously ignored. Complementary to SAS, HAR mitigates the energy overhead of on-device rendering through a new hardware accelerator that is specialized for projective transformation. We implement an EVR prototype on an Amazon AWS server instance and a NVIDA Jetson TX2 board combined with a Xilinx Zynq-7000 FPGA. Real system measurements show that EVR reduces the energy of VR rendering by up to 58%, which translates to up to 42% energy saving for VR devices.