Abstract
A line-scanning ophthalmoscope (LSO) is a retinal imaging technique that has the characteristics of high imaging resolution, wide field of view, and high imaging speed. However, the high-speed imaging with rather short exposure time inevitably reduces the signal intensity, and many factors, such as speckle noise and intraocular scatter, further degrade the signal-to-noise ratio (SNR) of retinal images. To effectively improve the image quality without increasing the LSO system’s complexity, the post-processing method of image super-resolution (SR) is adopted. In this paper, we propose a learning-based multi-frame retinal image SR method that directly learns an end-to-end mapping from low-resolution (LR) image sequences to high-resolution (HR) images. This network was validated on down-sampled and real LSO image sequences. We evaluated the method on a down-sampled dataset with the metrics of peak signal-to-noise ratio (PSNR), structural similarity (SSIM), and perceptual distance. Moreover, the power spectra and full width at half maximum (FWHM) were used as the no-reference image quality assessment (NR-IQA) algorithms to evaluate the reconstruction results of the real LSO image sequences. The experimental results indicate that the proposed method can significantly enhance the SNR of LSO images and efficiently improve the resolution of LSO retinal images, which has great practical significance for clinical diagnosis and analysis.
Highlights
Retinal imaging is one of the most common modalities of clinical practice in diagnosing retinal diseases
The test was conducted on the 16 line-scanning ophthalmoscope (LSO) retinal image sequences of our test set, with the criteria of peak signal-tonoise ratio (PSNR) and structural similarity (SSIM), calculated as follows: MSE =
In training the multi-frame image super-resolution GAN (MSRGAN) model, we introduced pixel-space loss (L1 loss) and feature-space loss (VGG loss) to help produce superresolution images of high perceptual quality
Summary
Retinal imaging is one of the most common modalities of clinical practice in diagnosing retinal diseases. The confocal scanning laser ophthalmoscope (CSLO) [1] is a confocal imaging technique that can produce high-resolution retinal images by two-dimensional (2D) scanning illumination and filtration of stray light through a confocal arranged pinhole. The 2D scanning of the CSLO results in an imaging speed usually lower than 20 Hz, such a low imaging speed will cause serious intra-frame jitter and severely. An improved line-scanning ophthalmoscope (LSO) technique [2] based on line-beam illumination and probing imaging is used for retinal imaging, which can greatly improve imaging speed, typically above 100 Hz. The most direct solution to increase imaging speed is to reduce the exposure time and control the exposure available. The signalto-noise ratio (SNR) of the LSO images decreases. Affected by speckle noise, intraocular scatter, and other factors, the quality of LSO images will be further degraded, and important fine morphological features will be further obscured, creating a challenge to any follow-up image analysis, such as retinal vessels segmentation and lesion
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
