Korean Journal of Optics and Photonics (한국광학회지)

Optical Society of Korea (한국광학회)
권호 표지
DOI QR코드

DOI QR Code

Speckle Noise Reduction and Image Quality Improvement in U-net-based Phase Holograms in BL-ASM

BL-ASM에서 U-net 기반 위상 홀로그램의 스펙클 노이즈 감소와 이미지 품질 향상

  • Oh-Seung Nam (School of Information and Communications Engineering, Chungbuk National University) ;
  • Ki-Chul Kwon (School of Information and Communications Engineering, Chungbuk National University) ;
  • Jong-Rae Jeong (Department of Information and Communication Engineering, Suwon Science College) ;
  • Kwon-Yeon Lee (Department of Electronic Engineering, Sunchon National University) ;
  • Nam Kim (School of Information and Communications Engineering, Chungbuk National University)
  • 남오승 (충북대학교 정보통신공학부) ;
  • 권기철 (충북대학교 정보통신공학부) ;
  • 정종래 (수원과학대학교 정보통신공학과) ;
  • 이권연 (국립순천대학교 전자공학과) ;
  • 김남 (충북대학교 정보통신공학부)
  • Received : 2023.07.05
  • Accepted : 2023.08.11
  • Published : 2023.10.25
Download PDF
⟨ Previous Next ⟩

Abstract

The band-limited angular spectrum method (BL-ASM) causes aliasing errors due to spatial frequency control problems. In this paper, a sampling interval adjustment technique for phase holograms and a technique for reducing speckle noise and improving image quality using a deep-learningbased U-net model are proposed. With the proposed technique, speckle noise is reduced by first calculating the sampling factor and controlling the spatial frequency by adjusting the sampling interval so that aliasing errors can be removed in a wide range of propagation. The next step is to improve the quality of the reconstructed image by learning the phase hologram to which the deep learning model is applied. In the S/W simulation of various sample images, it was confirmed that the peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) were improved by 5% and 0.14% on average, compared with the existing BL-ASM.

Band-limited angular spectrum method (BL-ASM)는 공간주파수 제어의 문제로 aliasing 오류가 발생한다. 본 논문에서는 위상 홀로그램에 대한 표본화 간격 조정 기법과 딥 러닝 기반의 U-net 모델을 사용한 스펙클 노이즈 감소 및 이미지 품질 향상 기법을 제안하였다. 제안한 기법에서는 넓은 전파 범위에서 aliasing 오류를 제거할 수 있도록 먼저 샘플링 팩터를 계산하여 표본화 간격 조절에 의한 공간주파수를 제어함으로써 스펙클 노이즈를 감소시킨다. 그 후 딥 러닝 모델을 적용한 위상 홀로그램을 학습시켜 복원 이미지의 품질을 향상시킨다. 다양한 샘플 이미지에 대한 S/W 시뮬레이션에서 기존의 BL-ASM과의 peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM)을 비교할 때 각각 평균 5%, 0.14% 정도 비율이 향상됨을 확인하였다.

Keywords

Acknowledgement

본 연구는 2023학년도 충북대학교 연구년제 사업의 연구비 지원과 정부(과학기술정보통신부)의 재원으로 정보통신기획평가원의 지원을 받아 수행된 연구임(IITP-2023-2020-0-01846, No. 2020-0-00929, No. 2021-0-00750).

References

  1. P. Chakravarthula, Y. Peng, J. Kollin, H. Fuchs, and F. Heide, "Wirtinger holography for near-eye displays," ACM Trans. Graph. 38, 213 (2019).
  2. Y. K. Kim, J. S. Lee, and Y. H. Won, "Low-noise high-efficiency double-phase hologram by multiplying a weight factor," Opt. Lett. 44, 3649-3652 (2019). https://doi.org/10.1364/OL.44.003649
  3. Y. Zhao, C.-X. Shi, K.-C. Kwon, Y.-L. Piao, M.-L. Piao, and N. Kim, "Fast calculation method of computer-generated hologram using a depth camera with point cloud gridding," Opt. Commun. 411, 166-169 (2018). https://doi.org/10.1016/j.optcom.2017.11.040
  4. L. Shi, B. Li, C. Kim, P. Kellnhofer, and W. Matusik, "Towards real-time photorealistic 3D holography with deep neural networks," Nature 591, 234-239 (2021). https://doi.org/10.1038/s41586-020-03152-0
  5. Y. Peng, S. Choi, N. Padmanaban, and G. Wetzstein, "Neural holography with camera-in-the-loop training," ACM Trans. Graph. 39, 185 (2020).
  6. J. Wu, K. Liu, X. Sui, and L. Cao, "High-speed computergenerated holography using an autoencoder-based deep neural network," Opt. Lett. 46, 2908-2911 (2021). https://doi.org/10.1364/OL.425485
  7. K. Matsushima and T. Shimobaba, "Band limited angular spectrum method for numerical simulation of free-space propagation in far and near fields," Opt. Express 17, 19662-19673 (2009). https://doi.org/10.1364/OE.17.019662
  8. J. Cha, H. Ban, S. Choi, J. Kim, and H. Kim, "A study on reconstruction performance of phase-only holograms with varying propagation distance," J. Broadcast Eng. 28, 3-20 (2023).
  9. J. Cho and Y. Kim, "DIML/CVL RGB-D dataset: 2M RGBD images of natural indoor and outdoor scenes," arXiv: 2110.11590v1 (2021).
  10. Y. Kim, B. Ham, C. Oh, and K. Sohn, "Structure selective depth super-resolution for RGB-D cameras," IEEE Trans. Image Process. 25, 5527-5538 (2016).
  11. S. Kim, D. Min, B. Ham, S. Kim, and K. Sohn, "Deep stereo confidence prediction for depth estimation," in Proc. 2017 IEEE International Conference on Image Processing-ICIP (Bejing, China, Sep. 17-20, 2017), pp. 992-996.
  12. Y. Kim, H. Jung, D. Min, and K. Sohn, "Deep monocular depth estimation via integration of global and local predictions," IEEE Trans. Image Process. 27, 4131-4143 (2018). https://doi.org/10.1109/TIP.2018.2836318
  13. J. Cho, D. Min, Y. Kim, and K. Sohn, "Deep monocular depth estimation leveraging a large-scale outdoor stereo dataset," Expert Syst. Appl. 178, 114877 (2021).