A Real-Time Learning-Based Super-Resolution System Using Direct Simple Functions

Abstract

This paper proposes a real-time super-resolution (SR) system. The proposed system performs a fast SR algorithm that generates a high-resolution image from a low-resolution image using direct regression functions. The system implemented on a Xilinx Virtex 7 field programmable gate array achieves output resolution of 3840 × 2160 (UHD) at 200 fps and 2000Mpixels/s throughput. Experimental results show that the proposed system provides high image quality for real-time applications.