Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Jan 14;11(2):831-849.
doi: 10.1364/BOE.380224. eCollection 2020 Feb 1.

Quality improvement of adaptive optics retinal images using conditional adversarial networks

Wanyue Li  1   2   3   4 , Guangxing Liu  3   4 , Yi He  2   3   5 , Jing Wang  1   2   3 , Wen Kong  1   2   3 , Guohua Shi  1   2   3   6   7
Affiliations

Quality improvement of adaptive optics retinal images using conditional adversarial networks

Wanyue Li et al. Biomed Opt Express. .

Abstract

The adaptive optics (AO) technique is widely used to compensate for ocular aberrations and improve imaging resolution. However, when affected by intraocular scatter, speckle noise, and other factors, the quality of the retinal image will be degraded. To effectively improve the image quality without increasing the imaging system's complexity, the post-processing method of image deblurring is adopted. In this study, we proposed a conditional adversarial network-based method for directly learning an end-to-end mapping between blurry and restored AO retinal images. The proposed model was validated on synthetically generated AO retinal images and real retinal images. The restoration results of synthetic images were evaluated with the metrics of peak signal-to-noise ratio (PSNR), structural similarity (SSIM), perceptual distance, and error rate of cone counting. Moreover, the blind image quality index (BIQI) was used as the no-reference image quality assessment (NR-IQA) algorithm to evaluate the restoration results on real AO retinal images. The experimental results indicate that the images restored by the proposed method have sharper quality and higher signal-to-noise ratio (SNR) when compared with other state-of-the-art methods, which has great practical significance for clinical research and analysis.

PubMed Disclaimer

Conflict of interest statement

The authors declare that there are no conflicts of interest related to this article.

Figures

Fig. 1.
Fig. 1.
Proposed model architecture. Our model is composed of a U-net based generator and a PatchGAN-based discriminator. (conv: convolutional layer; BN: batch normalization layer; ReLU: rectified linear unit; LReLU: leaky rectified linear unit; Tanh: TanHyperbolic function)
Fig. 2.
Fig. 2.
Deblurring results of synthetic AO retinal images imitating at different eccentricities and noises. (a1)-(a3) Ground truth; (b1) blurry images (0.3 mm eccentricity from the foveal center, and the Gaussian noise with a standard deviation of 0.03); (b2) blurry images (0.5 mm eccentricity from the foveal center, and the Gaussian noise with a standard deviation of 0.02); (b3) blurry images (1.0 mm eccentricity from the foveal center, and the Gaussian noise with a standard deviation of 0.05). Restored images from (c1), (c2), (c3) the ALM; (d1), (d2), (d3) the DeblurGAN method; (e1), (e2), (e3) the SRNdeblur method; and (f1), (f2), (f3) the proposed method. (Zooming-in the figure will provide a better look at the restoration quality).
Fig. 3.
Fig. 3.
Results of cone detection on synthetic retinal images. (a) Original blurry image (with small residual wavefront aberrations and Gaussian noise with a standard deviation of 0.01). Cone detection on (b) ground truth (1.5 mm eccentricity from the foveal center); (c) blurry image; image restored by (d) the ALM, (e) the DeblurGAN method, (f) the SRNdeblur method, and (g) the proposed method.
Fig. 4.
Fig. 4.
Results of cone detection on synthetic retinal images. (a) Original blurry image (with large residual wavefront aberrations and Gaussian noise with a standard deviation of 0.05). Cone detection on (b) ground truth (1.0 mm eccentricity from the foveal center); (c) blurry image; image restored by (d) the ALM, (e) the DeblurGAN method, (f) the SRNdeblur method, and (g) the proposed method.
Fig. 5.
Fig. 5.
Deblurring results of real retinal images captured by the AOSLO system [40]. (a) Original image (0.8 mm eccentricity from the foveal center). Restored images from (b) the ALM; (c) the DeblurGAN method; (d) the SRNdeblur method; and (e) the proposed method. (f) The corresponding image average power spectra. (Zooming-in the figure will provide a better look at the restoration quality).
Fig. 6.
Fig. 6.
Deblurring results of real retinal images captured by the AOSLO system [41]. (a) Original image (1.2 mm eccentricity from the foveal center). Restored images from (b) the ALM; (c) the DeblurGAN method; (d) the SRNdeblur method; and (e) the proposed method. (f) Corresponding image average power spectra. (Zooming-in the figure will provide a better look at the restoration quality).
Fig. 7.
Fig. 7.
Deblurring results of real retinal images captured by AOSLO system [41]. (a) Original image (1.8 mm eccentricity from the foveal center). Restored images from (b) the ALM; (c) the DeblurGAN method; (d) the SRNdeblur method; and (e) the proposed method. (f) The corresponding image average power spectra. (Zooming-in the figure will provide a better look at the restoration quality).
See this image and copyright information in PMC

References

    1. Liang J., Williams D. R., Miller D. T., “Supernormal vision and high-resolution retinal imaging through adaptive optics,” J. Opt. Soc. Am. A 14(11), 2884–2892 (1997).10.1364/JOSAA.14.002884 - DOI - PubMed
    1. Vargasmartin F., Prieto P. M., Artal P., “Correction of the aberrations in the human eye with a liquid-crystal spatial light modulator: limits to performance,” J. Opt. Soc. Am. A 15(9), 2552–2562 (1998).10.1364/JOSAA.15.002552 - DOI - PubMed
    1. Roorda A., Romeroborja F., Donnelly W. J., Queener H. M., Hebert T. J., Campbell M. C. W., “Adaptive optics scanning laser ophthalmoscopy,” Opt. Express 10(9), 405–412 (2002).10.1364/OE.10.000405 - DOI - PubMed
    1. Burns S. A., Marcos S., Elsner A. E., Bara S., “Contrast improvement of confocal retinal imaging by use of phase-correcting plates,” Opt. Lett. 27(6), 400–402 (2002).10.1364/OL.27.000400 - DOI - PMC - PubMed
    1. Lu J., Li H., Wei L., Shi G., Zhang Y., “Retina imaging in vivo with the adaptive optics confocal scanning laser ophthalmoscope,” in IEEE International Conference on Photonics (2009), 7519.

LinkOut - more resources