OSCR

Photorealistic 3D Holographic Display with Natural Defocus Effect.

Code ↔ Paper

8 matches between paragraphs of the paper and lines of its authors' code, computed by the harvester (lexical-v1). Click a colored paragraph or line to see its counterpart.

The 8 matches
  1. [1] § Methods › Simulation settings ↔ util.py, lines 1316–1375 · score 0.71 · anti aliasing double, Gaussian kernel, TensorHolography, phase, depths
  2. [2] § Methods › Simulation settings ↔ models/networks.py, lines 730–757 · score 0.71 · anti aliasing double, Gaussian kernel, network, phase, SLM
  3. [3] § Results › Simulation results of realistic defocus effect ↔ eval/eval_holo_rgb.py, lines 24–113 · score 0.71 · learned perceptual image, visual information fidelity, DISTS, VIF, LPIPS, patch
  4. [4] § Results › Simulation results of realistic defocus effect ↔ eval/eval_holo_rgb_d0.002.py, lines 24–110 · score 0.71 · learned perceptual image, visual information fidelity, DISTS, VIF, LPIPS, patch
  5. [5] § Methods › Simulation settings ↔ train/train_holo_new.py, lines 123–207 · score 0.60 · DeepFocus, pixel pitch, Adam, wavelengths, GPU, optimize
  6. [6] § Methods › Simulation settings ↔ train/train_holo_new_d0.002.py, lines 120–191 · score 0.60 · DeepFocus, pixel pitch, Adam, wavelengths, GPU, optimize
  7. [7] § Methods › The Analysis of Defocus Blur Perception Under Coherent and Incoherent Illumination ↔ FSIM.py, lines 127–270 · score 0.58 · polar coordinates, spread function, radius, zero
  8. [8] § Methods › The Analysis of Defocus Blur Perception Under Coherent and Incoherent Illumination ↔ FSIM.py, lines 127–270 · score 0.55 · polar coordinate, spread function, wavelength

Paper

Loaded from Europe PMC by your browser, not stored by OSCR: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The authors' code

Python · 371 lines · 13 KB · no license · 2 matches

  1. import math
  2. from typing import Tuple, Optional, Union
  4. import numpy as np
  5. import torch
  6. import torch.nn.functional as F
  9. def _to_tensor(img: Union[np.ndarray, torch.Tensor], device: Optional[torch.device] = None) -> torch.Tensor:
  10. """
  11. Convert input image to 2D or 3D torch tensor (H x W or H x W x 3) with dtype float32.
  12. Supports NumPy arrays and torch.Tensor. Values are assumed in range [0, 255].
  13. """
  14. if isinstance(img, np.ndarray):
  15. t = torch.from_numpy(img)
  16. elif isinstance(img, torch.Tensor):
  17. t = img
  18. else:
  19. raise TypeError("image must be a numpy.ndarray or torch.Tensor")
  21. if t.dtype not in (torch.float32, torch.float64):
  22. t = t.float()
  24. if device is not None:
  25. t = t.to(device)
  27. # Ensure 2D or 3D (H, W) or (H, W, 3)
  28. if t.ndim == 2:
  29. return t
  30. elif t.ndim == 3 and t.shape[2] == 3:
  31. return t
  32. elif t.ndim == 3 and t.shape[0] == 3:
  33. # Convert from C x H x W -> H x W x C
  34. return t.permute(1, 2, 0).contiguous()
  35. else:
  36. raise ValueError("Unsupported image shape. Expected 2D (H, W) or 3D with 3 channels (H, W, 3).")
  39. def _avg_downsample(x: torch.Tensor, stride: int) -> torch.Tensor:
  40. """
  41. Average downsample by factor stride using avg_pool2d.
  42. x: 2D tensor (H x W)
  43. """
  44. if stride <= 1:
  45. return x
  46. x4 = x.unsqueeze(0).unsqueeze(0) # 1 x 1 x H x W
  47. y = F.avg_pool2d(x4, kernel_size=stride, stride=stride)
  48. return y.squeeze(0).squeeze(0)
  51. def _conv2_same(x: torch.Tensor, kernel: torch.Tensor) -> torch.Tensor:
  52. """
  53. 2D convolution with 'same' output size using padding. x: 2D H x W, kernel: 2D kh x kw.
  54. """
  55. kh, kw = kernel.shape
  56. pad_h = kh // 2
  57. pad_w = kw // 2
  58. x4 = x.unsqueeze(0).unsqueeze(0) # N=1, C=1
  59. k4 = kernel.unsqueeze(0).unsqueeze(0)
  60. y = F.conv2d(x4, k4, padding=(pad_h, pad_w))
  61. return y.squeeze(0).squeeze(0)
  64. def _fftshift(x: torch.Tensor) -> torch.Tensor:
  65. """
  66. fftshift for 2D tensors.
  67. """
  68. h, w = x.shape
  69. return torch.roll(torch.roll(x, shifts=(h // 2), dims=0), shifts=(w // 2), dims=1)
  72. def _ifftshift(x: torch.Tensor) -> torch.Tensor:
  73. """
  74. ifftshift for 2D tensors.
  75. """
  76. h, w = x.shape
  77. return torch.roll(torch.roll(x, shifts=-(h // 2), dims=0), shifts=-(w // 2), dims=1)
  80. def _real_pow_lam(x: torch.Tensor, lam: float) -> torch.Tensor:
  81. """
  82. Real part of x^lam for real x and fractional lam.
  83. For negative x: Re(x^lam) = |x|^lam * cos(pi * lam)
  84. Avoids NaNs when x < 0 and lam is non-integer.
  85. """
  86. base = torch.clamp(torch.abs(x), min=1e-12)
  87. mag = base ** lam
  88. cos_term = math.cos(math.pi * lam)
  89. factor = torch.where(x >= 0, torch.ones_like(x), torch.full_like(x, cos_term))
  90. return mag * factor
  93. def lowpassfilter(sze: Tuple[int, int], cutoff: float, n: int, device: Optional[torch.device] = None) -> torch.Tensor:
  94. """
  95. Constructs a low-pass Butterworth filter (replicates MATLAB implementation).
  96. sze: (rows, cols)
  97. cutoff: 0..0.5
  98. n: integer >= 1
  99. Returns filter with frequency origin at corners (ifftshift applied).
  100. """
  101. rows, cols = int(sze[0]), int(sze[1])
  103. if cutoff < 0.0 or cutoff > 0.5:
  104. raise ValueError("cutoff frequency must be between 0 and 0.5")
  105. if int(n) != n or n < 1:
  106. raise ValueError("n must be an integer >= 1")
  108. # X/Y ranges normalized to +/- 0.5 (MATLAB-style handling of odd/even sizes)
  109. if cols % 2:
  110. xrange = torch.linspace(-(cols - 1) / 2.0, (cols - 1) / 2.0, cols, device=device) / (cols - 1)
  111. else:
  112. xrange = torch.linspace(-cols / 2.0, (cols / 2.0 - 1), cols, device=device) / cols
  114. if rows % 2:
  115. yrange = torch.linspace(-(rows - 1) / 2.0, (rows - 1) / 2.0, rows, device=device) / (rows - 1)
  116. else:
  117. yrange = torch.linspace(-rows / 2.0, (rows / 2.0 - 1), rows, device=device) / rows
  119. x, y = torch.meshgrid(yrange, xrange, indexing="ij")
  120. radius = torch.sqrt(x ** 2 + y ** 2)
  122. f = 1.0 / (1.0 + (radius / cutoff) ** (2 * n))
  123. f = _ifftshift(f)
  124. return f
  127. def phasecong2(
  128. im: torch.Tensor,
  129. nscale: int = 4,
  130. norient: int = 4,
  131. minWaveLength: float = 6.0,
  132. mult: float = 2.0,
  133. sigmaOnf: float = 0.55,
  134. dThetaOnSigma: float = 1.2,
  135. k: float = 2.0,
  136. epsilon: float = 1e-4,
  137. ) -> torch.Tensor:
  138. """
  139. Phase congruency map (replicates MATLAB's phasecong2 for FSIM).
  140. im: 2D torch.Tensor, dtype float32/64, expected intensity image
  141. Returns: ResultPC as 2D torch.Tensor
  142. """
  143. device = im.device
  144. rows, cols = im.shape
  145. imagefft = torch.fft.fft2(im)
  147. zero = torch.zeros(rows, cols, device=device)
  149. # Precompute polar coordinate grids with MATLAB-style center handling
  150. if cols % 2:
  151. xrange = torch.linspace(-(cols - 1) / 2.0, (cols - 1) / 2.0, cols, device=device) / (cols - 1)
  152. else:
  153. xrange = torch.linspace(-cols / 2.0, (cols / 2.0 - 1), cols, device=device) / cols
  155. if rows % 2:
  156. yrange = torch.linspace(-(rows - 1) / 2.0, (rows - 1) / 2.0, rows, device=device) / (rows - 1)
  157. else:
  158. yrange = torch.linspace(-rows / 2.0, (rows / 2.0 - 1), rows, device=device) / rows
  160. x, y = torch.meshgrid(yrange, xrange, indexing="ij")
  161. radius = torch.sqrt(x ** 2 + y ** 2)
  162. theta = torch.atan2(-y, x)
  163. radius = _ifftshift(radius)
  164. theta = _ifftshift(theta)
  165. radius[0, 0] = 1.0
  167. sintheta = torch.sin(theta)
  168. costheta = torch.cos(theta)
  170. # Low-pass filter
  171. lp = lowpassfilter((rows, cols), 0.45, 15, device=device)
  173. # Radial log-Gabor filters
  174. logGabor = []
  175. for s in range(nscale):
  176. wavelength = minWaveLength * (mult ** s)
  177. fo = 1.0 / wavelength
  178. logRad = torch.log(radius / fo)
  179. logG = torch.exp(-(logRad ** 2) / (2.0 * (math.log(sigmaOnf) ** 2)))
  180. logG = logG * lp
  181. logG[0, 0] = 0.0
  182. logGabor.append(logG)
  184. # Angular spread functions for orientations
  185. thetaSigma = math.pi / norient / dThetaOnSigma
  186. spread = []
  187. for o in range(norient):
  188. angl = (o) * math.pi / norient
  189. ds = sintheta * math.cos(angl) - costheta * math.sin(angl)
  190. dc = costheta * math.cos(angl) + sintheta * math.sin(angl)
  191. dtheta = torch.abs(torch.atan2(ds, dc))
  192. spread.append(torch.exp(-(dtheta ** 2) / (2.0 * thetaSigma ** 2)))
  194. EnergyAll = torch.zeros(rows, cols, device=device)
  195. AnAll = torch.zeros(rows, cols, device=device)
  197. # Pre-alloc for noise estimation filters
  198. ifftFilterArray = [torch.zeros(rows, cols, device=device) for _ in range(nscale)]
  200. for o in range(norient):
  201. sumE_ThisOrient = torch.zeros(rows, cols, device=device)
  202. sumO_ThisOrient = torch.zeros(rows, cols, device=device)
  203. sumAn_ThisOrient = torch.zeros(rows, cols, device=device)
  204. Energy = torch.zeros(rows, cols, device=device)
  206. maxAn = None
  207. EM_n = 0.0
  209. EO = []
  210. for s in range(nscale):
  211. filt = logGabor[s] * spread[o]
  212. ifftFilt = torch.real(torch.fft.ifft2(torch.complex(filt, torch.zeros_like(filt)))) * math.sqrt(rows * cols)
  213. ifftFilterArray[s] = ifftFilt
  214. # Convolve image in frequency domain: multiply FFT by filter
  215. eo = torch.fft.ifft2(imagefft * torch.complex(filt, torch.zeros_like(filt)))
  216. EO.append(eo)
  218. An = torch.abs(eo)
  219. sumAn_ThisOrient = sumAn_ThisOrient + An
  220. sumE_ThisOrient = sumE_ThisOrient + torch.real(eo)
  221. sumO_ThisOrient = sumO_ThisOrient + torch.imag(eo)
  223. if s == 0:
  224. EM_n = torch.sum(filt ** 2).item()
  225. maxAn = An.clone()
  226. else:
  227. maxAn = torch.maximum(maxAn, An)
  229. XEnergy = torch.sqrt(sumE_ThisOrient ** 2 + sumO_ThisOrient ** 2) + epsilon
  230. MeanE = sumE_ThisOrient / XEnergy
  231. MeanO = sumO_ThisOrient / XEnergy
  233. for s in range(nscale):
  234. E = torch.real(EO[s])
  235. O = torch.imag(EO[s])
  236. Energy = Energy + E * MeanE + O * MeanO - torch.abs(E * MeanO - O * MeanE)
  238. # Noise compensation
  239. eo0 = EO[0]
  240. E2n = torch.abs(eo0) ** 2
  241. medianE2n = torch.median(E2n.view(-1)).item()
  242. meanE2n = -medianE2n / math.log(0.5)
  243. noisePower = meanE2n / EM_n
  245. EstSumAn2 = torch.zeros(rows, cols, device=device)
  246. for s in range(nscale):
  247. EstSumAn2 = EstSumAn2 + ifftFilterArray[s] ** 2
  249. EstSumAiAj = torch.zeros(rows, cols, device=device)
  250. for si in range(nscale - 1):
  251. for sj in range(si + 1, nscale):
  252. EstSumAiAj = EstSumAiAj + ifftFilterArray[si] * ifftFilterArray[sj]
  254. sumEstSumAn2 = torch.sum(EstSumAn2).item()
  255. sumEstSumAiAj = torch.sum(EstSumAiAj).item()
  257. EstNoiseEnergy2 = 2.0 * noisePower * sumEstSumAn2 + 4.0 * noisePower * sumEstSumAiAj
  258. tau = math.sqrt(EstNoiseEnergy2 / 2.0)
  259. EstNoiseEnergy = tau * math.sqrt(math.pi / 2.0)
  260. EstNoiseEnergySigma = math.sqrt((2.0 - math.pi / 2.0) * tau * tau)
  262. T = EstNoiseEnergy + k * EstNoiseEnergySigma
  263. T = T / 1.7
  264. Energy = torch.maximum(Energy - T, zero)
  266. EnergyAll = EnergyAll + Energy
  267. AnAll = AnAll + sumAn_ThisOrient
  269. ResultPC = EnergyAll / (AnAll + 1e-12)
  270. return ResultPC
  273. def FeatureSIM(
  274. imageRef: Union[np.ndarray, torch.Tensor],
  275. imageDis: Union[np.ndarray, torch.Tensor],
  276. device: Optional[torch.device] = None,
  277. ) -> Tuple[float, float]:
  278. """
  279. Compute FSIM and FSIMc between two images (MATLAB FSIM.m port).
  280. Supports grayscale (H x W) and color (H x W x 3). Values expected 0..255.
  281. Returns: (FSIM, FSIMc)
  282. """
  283. ref = _to_tensor(imageRef, device=device)
  284. dis = _to_tensor(imageDis, device=device)
  286. rows, cols = ref.shape[0], ref.shape[1]
  287. is_color = (ref.ndim == 3 and ref.shape[2] == 3)
  289. # Build YIQ for color or use grayscale directly
  290. if is_color:
  291. R1, G1, B1 = ref[..., 0], ref[..., 1], ref[..., 2]
  292. R2, G2, B2 = dis[..., 0], dis[..., 1], dis[..., 2]
  294. Y1 = 0.299 * R1 + 0.587 * G1 + 0.114 * B1
  295. Y2 = 0.299 * R2 + 0.587 * G2 + 0.114 * B2
  296. I1 = 0.596 * R1 - 0.274 * G1 - 0.322 * B1
  297. I2 = 0.596 * R2 - 0.274 * G2 - 0.322 * B2
  298. Q1 = 0.211 * R1 - 0.523 * G1 + 0.312 * B1
  299. Q2 = 0.211 * R2 - 0.523 * G2 + 0.312 * B2
  300. else:
  301. Y1 = ref
  302. Y2 = dis
  303. # For grayscale, FSIMc == FSIM. We still define I/Q to keep code uniform.
  304. I1 = torch.ones(rows, cols, device=ref.device)
  305. I2 = torch.ones(rows, cols, device=ref.device)
  306. Q1 = torch.ones(rows, cols, device=ref.device)
  307. Q2 = torch.ones(rows, cols, device=ref.device)
  309. # Automatic downsampling
  310. minDimension = min(rows, cols)
  311. Fd = max(1, int(round(minDimension / 256.0)))
  313. Y1 = _avg_downsample(Y1, Fd)
  314. Y2 = _avg_downsample(Y2, Fd)
  315. I1 = _avg_downsample(I1, Fd)
  316. I2 = _avg_downsample(I2, Fd)
  317. Q1 = _avg_downsample(Q1, Fd)
  318. Q2 = _avg_downsample(Q2, Fd)
  320. # Phase congruency
  321. PC1 = phasecong2(Y1)
  322. PC2 = phasecong2(Y2)
  324. # Gradient maps (Sobel-like kernels used in FSIM)
  325. dx = torch.tensor([[3, 0, -3], [10, 0, -10], [3, 0, -3]], dtype=Y1.dtype, device=Y1.device) / 16.0
  326. dy = torch.tensor([[3, 10, 3], [0, 0, 0], [-3, -10, -3]], dtype=Y1.dtype, device=Y1.device) / 16.0
  328. IxY1 = _conv2_same(Y1, dx)
  329. IyY1 = _conv2_same(Y1, dy)
  330. gradientMap1 = torch.sqrt(IxY1 ** 2 + IyY1 ** 2)
  332. IxY2 = _conv2_same(Y2, dx)
  333. IyY2 = _conv2_same(Y2, dy)
  334. gradientMap2 = torch.sqrt(IxY2 ** 2 + IyY2 ** 2)
  336. # FSIM (luminance only)
  337. T1 = 0.85
  338. T2 = 160.0
  339. PCSimMatrix = (2.0 * PC1 * PC2 + T1) / (PC1 ** 2 + PC2 ** 2 + T1)
  340. gradientSimMatrix = (2.0 * gradientMap1 * gradientMap2 + T2) / (gradientMap1 ** 2 + gradientMap2 ** 2 + T2)
  341. PCm = torch.maximum(PC1, PC2)
  342. SimMatrix = gradientSimMatrix * PCSimMatrix * PCm
  343. denom = torch.sum(PCm)
  344. FSIM = (torch.sum(SimMatrix) / (denom + 1e-12)).item()
  346. # FSIMc (color)
  347. if is_color:
  348. T3 = 200.0
  349. T4 = 200.0
  350. ISimMatrix = (2.0 * I1 * I2 + T3) / (I1 ** 2 + I2 ** 2 + T3)
  351. QSimMatrix = (2.0 * Q1 * Q2 + T4) / (Q1 ** 2 + Q2 ** 2 + T4)
  353. lam = 0.03
  354. colorSim = _real_pow_lam(ISimMatrix * QSimMatrix, lam)
  355. SimMatrixC = gradientSimMatrix * PCSimMatrix * colorSim * PCm
  356. FSIMc = (torch.sum(SimMatrixC) / (denom + 1e-12)).item()
  357. else:
  358. FSIMc = FSIM
  360. return FSIM, FSIMc
  363. if __name__ == "__main__":
  364. # Example usage:
  365. # Load two images via numpy (H x W x 3), in range [0, 255]
  366. # img1 = np.asarray(...) # e.g., using imageio.imread
  367. # img2 = np.asarray(...)
  368. # fsim, fsimc = FeatureSIM(img1, img2)
  369. # print("FSIM:", fsim, "FSIMc:", fsimc)
  370. pass
  371. print("1")

FSIM.py at commit b1ad13d, no license · at the source

Overview

Authors: Mi Zhou1, Mu Ku Chen2, Fei Liu3, Mei Shen4, Lei Lei5, Chaoqun Ma1, Xueqian Wang1, Jian Song1, Haoqian Wang1, Kaichen Dong1, Chao Zuo6, Zihan Geng1,7
ORCID iDs: Mu Ku Chen, Xueqian Wang, Kaichen Dong, Chao Zuo, Zihan Geng
  1. Institute of Data and Information, Tsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen,Guangdong, China
  2. Department of Electrical Engineering, City University of Hong Kong,Hong Kong, China
  3. School of Optoelectronic Engineering, Xidian University,Xi’an, Shaanxi China
  4. College of Integrated Circuits, Shenzhen Polytechnic University, Shenzhen,Guangdong, China
  5. Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education and Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen,Guangdong, China
  6. School of Electronic and Optical Engineering, Nanjing University of Science and Technology,Nanjing, Jiangsu China
  7. Peng Cheng Laboratory, Shenzhen,Guangdong, China
Institutions: Tsinghua University (China); City University of Hong Kong (Hong Kong SAR China); Xidian University (China); Shenzhen Polytechnic University (China); Shenzhen University (China); Nanjing University of Science and Technology (China); Peng Cheng Laboratory (China)
Journal: Nature communications, volume 17, issue 1, article 6117
Dates: received 17 May 2025; accepted 23 April 2026; published online 6 May 2026
Type: Research article · Language: English
License: CC BY-NC-ND
Identifiers: DOI 10.1038/s41467-026-72736-7 · PMID 42091855 · PMCID PMC13357810 · OpenAlex W7160434489
Open access: gold, a free copy (OpenAlex)
Status: code verified
Methods: Physiology & signal measures
Keywords: Displays, Optical physics
Topic: Advanced Optical Imaging Technologies (Media Technology, Engineering), according to OpenAlex
Funding: National Science Foundation of China | National Natural Science Foundation of China-Yunnan Joint Fund (NSFC-Yunnan Joint Fund) (62305184); Guangdong Science and Technology Department (Science and Technology Department, Guangdong Province) (2023A1515012932); Shenzhen Science and Technology Innovation Commission (JCYJ20241202123919027); Research Grants Council, University Grants Committee (CRF Project: C5031-22G, and GRF Project: CityU11310522, CityU11300123); City University of Hong Kong (Project No. 9610628 and No. 7020142)
Citations: not cited yet (Europe PMC); 45 references in the paper

Abstract

The abstract is not reproduced here: the paper's license (CC BY-NC-ND) does not allow it. Read it in the paper, at the publisher or on Europe PMC.

Repository

Its files are read in the Code ↔ Paper reader above, with 8 matches between paragraphs and lines of code.

THUIntelligentOpticsLab/NaturalDefocus

License: none: the authors keep all their rights
State: the link answers, verified on 28 September 2026
Evidence: files inventoried
Commit: b1ad13dd65ce48657f52c2ea1aba970f291f4080, 19 March 2026
Languages: Python (37), Shell (4)
Size: 476 files, 41 scripts
Software Heritage: not archived
Found in: “Code availability”
Holds: README, tests
Not found: license file, CITATION.cff, environment file, continuous integration, documentation
Tools: PyTorch (30 files), NumPy (19 files), Matplotlib (13 files), Pillow (12 files), imageio (9 files), OpenCV (9 files), scikit-image (4 files), TensorFlow (2 files)
Availability: 1 check, the latest on 28 September 2026: the link answers
  • 28 September 2026: the link answers
43 files
At the source: github.com/THUIntelligentOpticsLab/NaturalDefocus

Code availability statement

The paper has a code availability statement. Its license (CC BY-NC-ND) does not allow reproducing it here; in short, from what the harvester recognized in it:

Read it in the paper: doi.org/10.1038/s41467-026-72736-7.

Tracing map

Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.

What the map holds:

  • 1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
  • 41 scripts, each with its path and the digest of its content;
  • 8 matches between paragraphs of the paper and lines of the code (method lexical-v1);
  • neither the text of the paper nor the code itself.

Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.

Data

No dataset and no data link were found in the paper.

Code and data availability statement

The paper has a code and data availability statement. Its license (CC BY-NC-ND) does not allow reproducing it here; in short, from what the harvester recognized in it:

Read it in the paper: doi.org/10.1038/s41467-026-72736-7.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 1, 28 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 12 authors, 2 keywords, 5 funders, 31 references.

Cite

This paper

Zhou, M., Chen, M. K., Liu, F., Shen, M., Lei, L., Ma, C., Wang, X., Song, J., Wang, H., Dong, K., Zuo, C., & Geng, Z. (2026). Photorealistic 3D Holographic Display with Natural Defocus Effect. Nature communications, 17(1), 6117. https://doi.org/10.1038/s41467-026-72736-7

BibTeX

@article{zhou2026photorealistic,
author = {Zhou, Mi and Chen, Mu Ku and Liu, Fei and Shen, Mei and Lei, Lei and Ma, Chaoqun and Wang, Xueqian and Song, Jian and Wang, Haoqian and Dong, Kaichen and Zuo, Chao and Geng, Zihan},
title = {{Photorealistic 3D Holographic Display with Natural Defocus Effect}},
journal = {Nature communications},
year = {2026},
month = may,
volume = {17},
number = {1},
pages = {6117},
publisher = {Nature Publishing Group},
issn = {2041-1723},
doi = {10.1038/s41467-026-72736-7},
url = {https://doi.org/10.1038/s41467-026-72736-7},
pmid = {42091855},
pmcid = {PMC13357810}
}

RIS

TY - JOUR
AU - Zhou, Mi
AU - Chen, Mu Ku
AU - Liu, Fei
AU - Shen, Mei
AU - Lei, Lei
AU - Ma, Chaoqun
AU - Wang, Xueqian
AU - Song, Jian
AU - Wang, Haoqian
AU - Dong, Kaichen
AU - Zuo, Chao
AU - Geng, Zihan
TI - Photorealistic 3D Holographic Display with Natural Defocus Effect
T2 - Nature communications
J2 - Nat Commun
PY - 2026
DA - 2026/05/06
VL - 17
IS - 1
SP - 6117
SN - 2041-1723
PB - Nature Publishing Group
DO - 10.1038/s41467-026-72736-7
UR - https://doi.org/10.1038/s41467-026-72736-7
LA - en
ER -

CSL-JSON

{
"id": "10.1038/s41467-026-72736-7",
"type": "article-journal",
"title": "Photorealistic 3D Holographic Display with Natural Defocus Effect",
"container-title": "Nature communications",
"author": [
{
"family": "Zhou",
"given": "Mi"
},
{
"family": "Chen",
"given": "Mu Ku"
},
{
"family": "Liu",
"given": "Fei"
},
{
"family": "Shen",
"given": "Mei"
},
{
"family": "Lei",
"given": "Lei"
},
{
"family": "Ma",
"given": "Chaoqun"
},
{
"family": "Wang",
"given": "Xueqian"
},
{
"family": "Song",
"given": "Jian"
},
{
"family": "Wang",
"given": "Haoqian"
},
{
"family": "Dong",
"given": "Kaichen"
},
{
"family": "Zuo",
"given": "Chao"
},
{
"family": "Geng",
"given": "Zihan"
}
],
"container-title-short": "Nat Commun",
"volume": "17",
"issue": "1",
"page": "6117",
"DOI": "10.1038/s41467-026-72736-7",
"PMID": "42091855",
"PMCID": "PMC13357810",
"ISSN": "2041-1723",
"publisher": "Nature Publishing Group",
"URL": "https://doi.org/10.1038/s41467-026-72736-7",
"language": "en",
"issued": {
"date-parts": [
[
2026,
5,
6
]
]
}
}

The tracing map gets a citation of its own once an author has validated it and it has a DOI.

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.1371/journal.pcbi.1014571 [code]
SynAPSeg: A novel dataset and image analysis framework for deep learning-based synapse detection and quantification.
Journal: PLoS computational biology
In common: imageio, TensorFlow, OpenCV, 5 other tools
[2] doi:10.1364/boe.605322 [code]
Generalized plaque digitization framework for multi-dimensional mesoscopic images.
Journal: Biomedical optics express
In common: imageio, TensorFlow, OpenCV, 5 other tools
[3] doi:10.1016/j.isci.2026.116168 [code]
See the small lesions: Frequency-guided spatial debiasing GAN for multimodal medical image fusion.
Journal: iScience
In common: imageio, TensorFlow, OpenCV, 5 other tools
[4] doi: [code]
Real-time closed-loop feedback system for mouse mesoscale cortical signal and movement control
Journal: eLife
In common: imageio, TensorFlow, OpenCV, 4 other tools
[5] doi:10.1002/epi.70296 [code]
Fully automated three-dimensional deep learning-based magnetic resonance imaging segmentation of brain cavities in epilepsy surgery.
Journal: Epilepsia
In common: imageio, TensorFlow, OpenCV, 4 other tools
[6] doi:10.1186/s12880-026-02481-2 [code]
Deep learning-based neuroanatomical profiling reveals population-specific brain changes in multiple sclerosis: a large-scale Middle Eastern study.
Journal: BMC medical imaging
In common: imageio, TensorFlow, OpenCV, 4 other tools
[7] doi:10.1016/j.xcrm.2026.102766 [code]
A longitudinal single-cell and spatial multiomic atlas of pediatric high-grade glioma.
Journal: Cell reports. Medicine
In common: imageio, TensorFlow, OpenCV, 4 other tools
[8] doi:10.1038/s41467-026-74358-5 [code]
Brain-inspired spatial intelligence for embodied agents.
Journal: Nature communications
In common: imageio, TensorFlow, OpenCV, 4 other tools
[9] doi:10.1038/s41598-026-43529-1 [code]
A spiking neural network inspired by neuroscience and psychology for Western mode- and key-conditioned music learning and composition.
Journal: Scientific reports
In common: imageio, TensorFlow, OpenCV, 4 other tools
[10] doi:10.1162/imag.a.1326 [code]
RAVEN: Robust, generalizable, multi-resolution structural MRI upsampling using autoencoders.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: imageio, OpenCV, scikit-image, 4 other tools

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion, reproductions, activity

Discussion: questions and error reports about this paper and its code, from signed-in readers and its authors. It opens with sign-in.

Reproductions: reports from readers who ran the authors' code: what they reproduced, with which environment, commit and data. It opens with sign-in.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.