FuseSR: Super Resolution for Real-time Rendering through Efficient Multi-resolution Fusion

TL;DR

FuseSR introduces a real-time super-resolution technique for high-resolution rendering that leverages auxiliary G-Buffers to produce high-quality, temporally consistent images efficiently, significantly reducing rendering overhead.

Contribution

The paper presents H-Net, a novel architecture that fuses multi-resolution features using G-Buffers for efficient super-resolution in real-time rendering, outperforming existing methods.

Findings

Achieves real-time 4x4 and 8x8 super-resolution at 4K resolution.

Produces temporally consistent high-quality reconstructions.

Reduces rendering overhead significantly compared to prior approaches.

Abstract

The workload of real-time rendering is steeply increasing as the demand for high resolution, high refresh rates, and high realism rises, overwhelming most graphics cards. To mitigate this problem, one of the most popular solutions is to render images at a low resolution to reduce rendering overhead, and then manage to accurately upsample the low-resolution rendered image to the target resolution, a.k.a. super-resolution techniques. Most existing methods focus on exploiting information from low-resolution inputs, such as historical frames. The absence of high frequency details in those LR inputs makes them hard to recover fine details in their high-resolution predictions. In this paper, we propose an efficient and effective super-resolution method that predicts high-quality upsampled reconstructions utilizing low-cost high-resolution auxiliary G-Buffers as additional input. With LR images and HR G-buffers as input, the network requires to align and fuse features at multi resolution levels. We introduce an efficient and effective H-Net architecture to solve this problem and significantly reduce rendering overhead without noticeable quality deterioration. Experiments show that our method is able to produce temporally consistent reconstructions in $4 \times 4$ and even challenging $8 \times 8$ upsampling cases at 4K resolution with real-time performance, with substantially improved quality and significant performance boost compared to existing works.