RLE Recent Papers

FAST: A Framework to Accelerate Super-Resolution Processing on Compressed Videos

Zhengdong Zhang, Vivienne Sze

DOI: 10.1109/CVPRW.2017.138

Abstract:

State-of-the-art super-resolution (SR) algorithms require significant computational resources to achieve real-time throughput (e.g., 60Mpixels/s for HD video). This paper in-troduces FAST (Free Adaptive Super-resolution via Trans-fer), a framework to accelerate any SR algorithm applied to compressed videos. FAST exploits the temporal corre-lation between adjacent frames such that SR is only ap-plied to a subset of frames; SR pixels are then transferred to the other frames. The transferring process has negli-gible computation cost as it uses information already em-bedded in the compressed video (e.g., motion vectors and residual). Adaptive processing is used to retain accuracy when the temporal correlation is not present (e.g., occlu-sions). FAST accelerates state-of-the-art SR algorithms by up to 15× with a visual quality loss of 0.2dB. FAST is an important step towards real-time SR algorithms for ultra-HD displays and energy constrained devices (e.g., phones and tablets).