High-speed reconstruction for ultra-low resolution faces

In this paper, a learning-based high-speed reconstruction system for ultra-low resolution faces is implemented 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 memories 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 reconstruction for ultra-low resolution faces. The system can reconstruct faces using 8x6, 16x12, and 32x24 sized images, with 4x4, 8x8, or 16x16 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 maximum of 7900 times faster than a pure software implementation using C.