首页> 外文会议>International Conference on Practical Holography: Displays, Materials, and Applications >Acceleration of phase-only hologram generation by sub-image preiteration
【24h】

Acceleration of phase-only hologram generation by sub-image preiteration

机译:通过子图像预夹加速相位全息图产生

获取原文

摘要

Holograms are a promising technology for realizing a natural three-dimensional display. While the holographic pattern records both the phase and intensity information of an object, the spatial light modulator, which is a device for reproducing it, is capable of modulating either phase or amplitude. Therefore, in order to realize a perfect holographic image, a complex hologram must be converted into a phase-only hologram. Currently, SLM resolution is 4k or less, but resolution needs to be increased to achieve better quality holograms. This will increase the computational complexity of the Fourier transform, which is needed for each iterate. Therefore, a way to create a phase-only-hologram with a minimum number of iterations is needed. It is known that in the iterative method, the initial predicted value greatly affects the result. In this study, we propose a method to make initial guess to obtain more accurate phase only hologram with fewer iteration times. To obtain the initial guess, the desired image was divided into several parts, and the phase-only hologram corresponding to each was obtained and added again. We used hybrid-input-output iteration to maximize the effect of pre-iterated guess. When iterates the same number of times, the peak signal-to-noise ratio (PSNR) value obtained by using proposed pre-iterated guess was higher than using the random initial guess. In other words, the desired PSNR value can be reached with a smaller number of iterations by using proposed method. After 15 iterations, proposed initial guess have ~0.3dB higher PSNR value, which means one or two times of reduced iterations.
机译:全息图是实现自然三维显示器的有希望的技术。虽然全息图案记录对象的相位和强度信息,但是空间光调制器,其是用于再现它的设备,能够调制双相或幅度。因此,为了实现完美的全息图像,必须将复杂的全息图转换成仅相位全息图。目前,SLM分辨率为4K或更小,但需要增加解决以实现更高的质量全息图。这将提高傅里叶变换的计算复杂性,这是每个迭代所需的变换。因此,需要一种使用最小迭代次数创建相位全息图的方法。众所周知,在迭代方法中,初始预测值极大地影响结果。在本研究中,我们提出了一种方法来使初始猜测能够使用较少的迭代时间获得更准确的阶段全息图。为了获得初始猜测,所需的图像被分成几个部分,并且获得了对应于每个的相位​​全息图并再次添加。我们使用混合输入输出迭代来最大化预迭代猜测的效果。当迭代相同的次数时,通过使用所提出的预迭代猜测获得的峰值信噪比(PSNR)值高于使用随机初始猜测。换句话说,通过使用所提出的方法可以通过较少数量的迭代来达到所需的PSNR值。在15次迭代之后,提出的初始猜测具有〜0.3dB的PSNR值,这意味着减少迭代的一倍或两倍。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号