超低分辨率人脸图像高速重建方法

High-speed reconstruction for ultra-low resolution faces

文中采用软硬件相结合的方法,实现了一种基于学习的超低分辨率人脸图像的高速重建系统.硬件系统的工作频率为60MHz,采用了FPGA组成并行处理单元,采用多内存形成并行数据,高速实现了计算复杂度高的相似度计算、相似度排序,从而解决了基于学习的超低分辨率人脸图像重建的处理速度问题.文中实现了8×6,16×12,32×24三个级别的人脸图像尺寸的重建系统,重建倍率分别为4×4倍、8×8倍、16×16倍,取得了好的视觉效果和低的RMS误差率,在速度上与C相比,最大可达到7900多倍的加速比,在保证重建质量的同时在处理速度上也有显著的提高.

In this paper, a learning-based high-speed reconstruction system for ultra-low resolution faces is im- plemented using a software/hardware co-design paradigm. The hardware component working at 60 MHz contains a field programmable gate array, which is reconfigured to contain parallel processing units, and multiple memo- ries to create parallel data. The hardware component effectively handles generating and sorting computationally intensive similarity metrics. This solves the processing speed problem in learning-based super-resolution recon- struction for ultra-low resolution faces. The system can reconstruct faces using 8 × 6, 16 × 12, and 32 × 24 sized images, with 4 ×4, 8 × 8, or 16 × 16 times magnification. The experimental results verify the effectiveness of our system in terms of both visual effect and low root mean square errors. The processing speed can be improved up to a maxinmm of 7900 times faster than a pure software implementation using C.