LIU Yi-peng, ZENG Hong-xiang, WANG Hai-xia, et al. Parallel design for sweat pore recognition algorithm based on embedded GPU[J]. 2018, (5): 76-79. DOI: 10.13873/j.1000-9787(2018)05-0076-04.
基于嵌入式GPU的汗孔识别算法并行设计
摘要
基于NVIDIA Jetson Tx2平台
结合Open CV计算机视觉库与计算统一设备架构(CUDA)程序设计
对汗孔特征提取与匹配算法实现了并行设计。实验结果表明:并行设计算法能够实现最多180倍的加速
推动指纹匹配算法在嵌入式系统领域的应用。
Abstract
Parallel design of extraction of sweat pore characteristics and matching algorithm is implemented by combining Open CV computer vision library with compute unified device architecture( CUDA) program design based on NVIDIA Jetson Tx2 platform.Experimental results show that parallel design algorithm can achieve speed up to at most 180 times faster
propel application of fingerprint matching algorithm in embedded systems.