CAJ | 학술논문

在ECT系统中使用支持向量机处理采集到的数据集,当数据规模非常大时训练速度缓慢,支持向量机在处理ECT系统中采集到的大规模数据集时训练速度缓慢.针对该问题提出了一种适应于硬件实现的基于SVM的串行计算-并行传输模式,并用硬件描述语言在现场可编程逻辑门阵列(FPGA)上实现,同时给出了硬件系统的工作过程及总体结构.ECT图像重建实验结果表明,同软件实现相比,不仅提高了系统图像重建速度,还能保持较高的分类精度,并且表明FPGA在图像重建方面满足实时性要求,适用于图像重建系统.
재ECT계통중사용지지향량궤처리채집도적수거집,당수거규모비상대시훈련속도완만,지지향량궤재처리ECT계통중채집도적대규모수거집시훈련속도완만.침대해문제제출료일충괄응우경건실현적기우SVM적천행계산-병행전수모식,병용경건묘술어언재현장가편정라집문진렬(FPGA)상실현,동시급출료경건계통적공작과정급총체결구.ECT도상중건실험결과표명,동연건실현상비,불부제고료계통도상중건속도,환능보지교고적분류정도,병차표명FPGA재도상중건방면만족실시성요구,괄용우도상중건계통.
Support vector machine (SVM) has slow training speed to deal with large amount of data in ECT system. In view of these shortcomings, an improved serial calculation-parallel transmission mode that based on SVM which is adapted to realize on hardware is proposed, and the mode on FPGA is realized using hardware description language. The work process of system and overall structure are presented. The ECT image reconstruction experimental results show that the hardware implementation of the algorithm reduces the running time efficiently, meanwhile not decreasing the classification accuracy. In the aspect of speed, it is more quickly using FPGA to realize the algorithm than using software, meanwhile FPGA can meet the request of real time and is fit for the system of image reconstruction.