Lee Yong H, Kim Kwan W, Kim Soon S
Department of Computer Engineering, Halla University, Wonju, Gangwon-Do, Korea.
Stud Health Technol Inform. 2013;192:987.
This paper presents a method to improve the contrast of digital X-ray image using CUDA program on a GPU. The histogram is commonly used to get the statistical distribution of the contrast in image processing. To increase the visibility of the image in real time, we use the histogram stretching function. It is difficult to implement the function on a GPU because the CUDA program is due to handle the complex process to transfer the source data and the processed results between the memory of GPU and the host system. As a result, we show to operate the histogram stretching function quickly on GPU by the CUDA program.
本文提出了一种在图形处理器(GPU)上使用统一计算设备架构(CUDA)程序来提高数字X射线图像对比度的方法。在图像处理中,直方图通常用于获取对比度的统计分布。为了实时提高图像的可视性,我们使用直方图拉伸函数。由于CUDA程序需要处理在GPU内存和主机系统之间传输源数据和处理结果的复杂过程,因此在GPU上实现该函数很困难。结果,我们展示了如何通过CUDA程序在GPU上快速运行直方图拉伸函数。