Research

Jump to: Review

17 pages, 4172 KiB

Open AccessArticle

Multi-Scale Learning with Sparse Residual Network for Explainable Multi-Disease Diagnosis in OCT Images

by Phuoc-Nguyen Bui, Duc-Tai Le, Junghyun Bum, Seongho Kim, Su Jeong Song and Hyunseung Choo

Bioengineering 2023, 10(11), 1249; https://doi.org/10.3390/bioengineering10111249 - 26 Oct 2023

Viewed by 1022

In recent decades, medical imaging techniques have revolutionized the field of disease diagnosis, enabling healthcare professionals to noninvasively observe the internal structures of the human body. Among these techniques, optical coherence tomography (OCT) has emerged as a powerful and versatile tool that allows [...] Read more.

In recent decades, medical imaging techniques have revolutionized the field of disease diagnosis, enabling healthcare professionals to noninvasively observe the internal structures of the human body. Among these techniques, optical coherence tomography (OCT) has emerged as a powerful and versatile tool that allows high-resolution, non-invasive, and real-time imaging of biological tissues. Deep learning algorithms have been successfully employed to detect and classify various retinal diseases in OCT images, enabling early diagnosis and treatment planning. However, existing deep learning algorithms are primarily designed for single-disease diagnosis, which limits their practical application in clinical settings where OCT images often contain symptoms of multiple diseases. In this paper, we propose an effective approach for multi-disease diagnosis in OCT images using a multi-scale learning (MSL) method and a sparse residual network (SRN). Specifically, the MSL method extracts and fuses useful features from images of different sizes to enhance the discriminative capability of a classifier and make the disease predictions interpretable. The SRN is a minimal residual network, where convolutional layers with large kernel sizes are replaced with multiple convolutional layers that have smaller kernel sizes, thereby reducing model complexity while achieving a performance similar to that of existing convolutional neural networks. The proposed multi-scale sparse residual network significantly outperforms existing methods, exhibiting 97.40% accuracy, 95.38% sensitivity, and 98.25% specificity. Experimental results show the potential of our method to improve explainable diagnosis systems for various eye diseases via visual discrimination. Full article

(This article belongs to the Special Issue Biomedical Imaging and Analysis of the Eye)

► Show Figures

Figure 1

21 pages, 30194 KiB

Open AccessArticle

Self-FI: Self-Supervised Learning for Disease Diagnosis in Fundus Images

by Toan Duc Nguyen, Duc-Tai Le, Junghyun Bum, Seongho Kim, Su Jeong Song and Hyunseung Choo

Bioengineering 2023, 10(9), 1089; https://doi.org/10.3390/bioengineering10091089 - 16 Sep 2023

Viewed by 1078

Abstract

Self-supervised learning has been successful in computer vision, and its application to medical imaging has shown great promise. This study proposes a novel self-supervised learning method for medical image classification, specifically targeting ultra-wide-field fundus images (UFI). The proposed method utilizes contrastive learning to [...] Read more.

Self-supervised learning has been successful in computer vision, and its application to medical imaging has shown great promise. This study proposes a novel self-supervised learning method for medical image classification, specifically targeting ultra-wide-field fundus images (UFI). The proposed method utilizes contrastive learning to pre-train a deep learning model and then fine-tune it with a small set of labeled images. This approach reduces the reliance on labeled data, which is often limited and costly to obtain, and has the potential to improve disease detection in UFI. This method employs two contrastive learning techniques, namely bi-lateral contrastive learning and multi-modality pre-training, to form positive pairs using the data correlation. Bi-lateral learning fuses multiple views of the same patient’s images, and multi-modality pre-training leverages the complementary information between UFI and conventional fundus images (CFI) to form positive pairs. The results show that the proposed contrastive learning method achieves state-of-the-art performance with an area under the receiver operating characteristic curve (AUC) score of 86.96, outperforming other approaches. The findings suggest that self-supervised learning is a promising direction for medical image analysis, with potential applications in various clinical settings. Full article

(This article belongs to the Special Issue Biomedical Imaging and Analysis of the Eye)

► Show Figures

Graphical abstract

17 pages, 15651 KiB

Open AccessArticle

Discriminative-Region Multi-Label Classification of Ultra-Widefield Fundus Images

by Van-Nguyen Pham, Duc-Tai Le, Junghyun Bum, Seong Ho Kim, Su Jeong Song and Hyunseung Choo

Bioengineering 2023, 10(9), 1048; https://doi.org/10.3390/bioengineering10091048 - 06 Sep 2023

Viewed by 957

Abstract

Ultra-widefield fundus image (UFI) has become a crucial tool for ophthalmologists in diagnosing ocular diseases because of its ability to capture a wide field of the retina. Nevertheless, detecting and classifying multiple diseases within this imaging modality continues to pose a significant challenge [...] Read more.

Ultra-widefield fundus image (UFI) has become a crucial tool for ophthalmologists in diagnosing ocular diseases because of its ability to capture a wide field of the retina. Nevertheless, detecting and classifying multiple diseases within this imaging modality continues to pose a significant challenge for ophthalmologists. An automated disease classification system for UFI can support ophthalmologists in making faster and more precise diagnoses. However, existing works for UFI classification often focus on a single disease or assume each image only contains one disease when tackling multi-disease issues. Furthermore, the distinctive characteristics of each disease are typically not utilized to improve the performance of the classification systems. To address these limitations, we propose a novel approach that leverages disease-specific regions of interest for the multi-label classification of UFI. Our method uses three regions, including the optic disc area, the macula area, and the entire UFI, which serve as the most informative regions for diagnosing one or multiple ocular diseases. Experimental results on a dataset comprising 5930 UFIs with six common ocular diseases showcase that our proposed approach attains exceptional performance, with the area under the receiver operating characteristic curve scores for each class spanning from 95.07% to 99.14%. These results not only surpass existing state-of-the-art methods but also exhibit significant enhancements, with improvements of up to 5.29%. These results demonstrate the potential of our method to provide ophthalmologists with valuable information for early and accurate diagnosis of ocular diseases, ultimately leading to improved patient outcomes. Full article

(This article belongs to the Special Issue Biomedical Imaging and Analysis of the Eye)

► Show Figures

Graphical abstract

12 pages, 1541 KiB

Open AccessArticle

Localized Refractive Changes Induced by Symmetric and Progressive Asymmetric Intracorneal Ring Segments Assessed with a 3D Finite-Element Model

by Gonzalo García de Oteyza, Juan Álvarez de Toledo, Rafael I. Barraquer and Sabine Kling

Bioengineering 2023, 10(9), 1014; https://doi.org/10.3390/bioengineering10091014 - 27 Aug 2023

Viewed by 899

Abstract

To build a representative 3D finite element model (FEM) for intracorneal ring segment (ICRS) implantation and to investigate localized optical changes induced by different ICRS geometries, a hyperelastic shell FEM was developed to compare the effect of symmetric and progressive asymmetric ICRS designs [...] Read more.

To build a representative 3D finite element model (FEM) for intracorneal ring segment (ICRS) implantation and to investigate localized optical changes induced by different ICRS geometries, a hyperelastic shell FEM was developed to compare the effect of symmetric and progressive asymmetric ICRS designs in a generic healthy and asymmetric keratoconic (KC) cornea. The resulting deformed geometry was assessed in terms of average curvature via a biconic fit, sagittal curvature (K), and optical aberrations via Zernike polynomials. The sagittal curvature map showed a locally restricted flattening interior to the ring (Kmax −11 to −25 dpt) and, in the KC cornea, an additional local steepening on the opposite half of the cornea (Kmax up to +1.9 dpt). Considering the optical aberrations present in the model of the KC cornea, the progressive ICRS corrected vertical coma (−3.42 vs. −3.13 µm); horizontal coma (−0.67 vs. 0.36 µm); and defocus (2.90 vs. 2.75 µm), oblique trefoil (−0.54 vs. −0.08 µm), and oblique secondary astigmatism (0.48 vs. −0.09 µm) aberrations stronger than the symmetric ICRS. Customized ICRS designs inspired by the underlying KC phenotype have the potential to achieve more tailored refractive corrections, particularly in asymmetric keratoconus patterns. Full article

(This article belongs to the Special Issue Biomedical Imaging and Analysis of the Eye)

► Show Figures

Figure 1

21 pages, 12323 KiB

Open AccessArticle

OculusGraphy: Signal Analysis of the Electroretinogram in a Rabbit Model of Endophthalmitis Using Discrete and Continuous Wavelet Transforms

by Aleksei Zhdanov, Paul Constable, Sultan Mohammad Manjur, Anton Dolganov, Hugo F. Posada-Quintero and Aleksander Lizunov

Bioengineering 2023, 10(6), 708; https://doi.org/10.3390/bioengineering10060708 - 11 Jun 2023

Cited by 3 | Viewed by 1215

Abstract

Background: The electroretinogram is a clinical test used to assess the function of the photoreceptors and retinal circuits of various cells in the eye, with the recorded waveform being the result of the summated response of neural generators across the retina. Methods: The [...] Read more.

Background: The electroretinogram is a clinical test used to assess the function of the photoreceptors and retinal circuits of various cells in the eye, with the recorded waveform being the result of the summated response of neural generators across the retina. Methods: The present investigation involved an analysis of the electroretinogram waveform in both the time and time–frequency domains through the utilization of the discrete wavelet transform and continuous wavelet transform techniques. The primary aim of this study was to monitor and evaluate the effects of treatment in a New Zealand rabbit model of endophthalmitis via electroretinogram waveform analysis and to compare these with normal human electroretinograms. Results: The wavelet scalograms were analyzed using various mother wavelets, including the Daubechies, Ricker, Wavelet Biorthogonal 3.1 (bior3.1), Morlet, Haar, and Gaussian wavelets. Distinctive variances were identified in the wavelet scalograms between rabbit and human electroretinograms. The wavelet scalograms in the rabbit model of endophthalmitis showed recovery with treatment in parallel with the time-domain features. Conclusions: The study compared adult, child, and rabbit electroretinogram responses using DWT and CWT, finding that adult signals had higher power than child signals, and that rabbit signals showed differences in the a-wave and b-wave depending on the type of response tested, while the Haar wavelet was found to be superior in visualizing frequency components in electrophysiological signals for following the treatment of endophthalmitis and may give additional outcome measures for the management of retinal disease. Full article

(This article belongs to the Special Issue Biomedical Imaging and Analysis of the Eye)

► Show Figures

Figure 1

11 pages, 1841 KiB

Open AccessArticle

Non-Invasive Hybrid Ultrasound Stimulation of Visual Cortex In Vivo

by Chen Gong, Runze Li, Gengxi Lu, Jie Ji, Yushun Zeng, Jiawen Chen, Chifeng Chang, Junhang Zhang, Lily Xia, Deepthi S. Rajendran Nair, Biju B. Thomas, Brian J. Song, Mark S. Humayun and Qifa Zhou

Bioengineering 2023, 10(5), 577; https://doi.org/10.3390/bioengineering10050577 - 10 May 2023

Cited by 4 | Viewed by 2025

Abstract

The optic nerve is the second cranial nerve (CN II) that connects and transmits visual information between the retina and the brain. Severe damage to the optic nerve often leads to distorted vision, vision loss, and even blindness. Such damage can be caused [...] Read more.

The optic nerve is the second cranial nerve (CN II) that connects and transmits visual information between the retina and the brain. Severe damage to the optic nerve often leads to distorted vision, vision loss, and even blindness. Such damage can be caused by various types of degenerative diseases, such as glaucoma and traumatic optic neuropathy, and result in an impaired visual pathway. To date, researchers have not found a viable therapeutic method to restore the impaired visual pathway; however, in this paper, a newly synthesized model is proposed to bypass the damaged portion of the visual pathway and set up a direct connection between a stimulated visual input and the visual cortex (VC) using Low-frequency Ring-transducer Ultrasound Stimulation (LRUS). In this study, by utilizing and integrating various advanced ultrasonic and neurological technologies, the following advantages are achieved by the proposed LRUS model: 1. This is a non-invasive procedure that uses enhanced sound field intensity to overcome the loss of ultrasound signal due to the blockage of the skull. 2. The simulated visual signal generated by LRUS in the visual-cortex-elicited neuronal response in the visual cortex is comparable to light stimulation of the retina. The result was confirmed by a combination of real-time electrophysiology and fiber photometry. 3. VC showed a faster response rate under LRUS than light stimulation through the retina. These results suggest a potential non-invasive therapeutic method for restoring vision in optic-nerve-impaired patients using ultrasound stimulation (US). Full article

(This article belongs to the Special Issue Biomedical Imaging and Analysis of the Eye)

► Show Figures

Figure 1

12 pages, 3662 KiB

Open AccessArticle

Reliability of Retinal Layer Annotation with a Novel, High-Resolution Optical Coherence Tomography Device: A Comparative Study

by Leon von der Emde, Marlene Saßmannshausen, Olivier Morelle, Geena Rennen, Frank G. Holz, Maximilian W. M. Wintergerst and Thomas Ach

Bioengineering 2023, 10(4), 438; https://doi.org/10.3390/bioengineering10040438 - 31 Mar 2023

Cited by 4 | Viewed by 3393

Abstract

Optical coherence tomography (OCT) enables in vivo diagnostics of individual retinal layers in the living human eye. However, improved imaging resolution could aid diagnosis and monitoring of retinal diseases and identify potential new imaging biomarkers. The investigational high-resolution OCT platform (High-Res OCT; 853 [...] Read more.

Optical coherence tomography (OCT) enables in vivo diagnostics of individual retinal layers in the living human eye. However, improved imaging resolution could aid diagnosis and monitoring of retinal diseases and identify potential new imaging biomarkers. The investigational high-resolution OCT platform (High-Res OCT; 853 nm central wavelength, 3 µm axial-resolution) has an improved axial resolution by shifting the central wavelength and increasing the light source bandwidth compared to a conventional OCT device (880 nm central wavelength, 7 µm axial-resolution). To assess the possible benefit of a higher resolution, we compared the retest reliability of retinal layer annotation from conventional and High-Res OCT, evaluated the use of High-Res OCT in patients with age-related macular degeneration (AMD), and assessed differences of both devices on subjective image quality. Thirty eyes of 30 patients with early/intermediate AMD (iAMD; mean age 75 ± 8 years) and 30 eyes of 30 age-similar subjects without macular changes (62 ± 17 years) underwent identical OCT imaging on both devices. Inter- and intra-reader reliability were analyzed for manual retinal layer annotation using EyeLab. Central OCT B-scans were graded for image quality by two graders and a mean-opinion-score (MOS) was formed and evaluated. Inter- and intra-reader reliability were higher for High-Res OCT (greatest benefit for inter-reader reliability: ganglion cell layer; for intra-reader reliability: retinal nerve fiber layer). High-Res OCT was significantly associated with an improved MOS (MOS 9/8, Z-value = 5.4, p < 0.01) mainly due to improved subjective resolution (9/7, Z-Value 6.2, p < 0.01). The retinal pigment epithelium drusen complex showed a trend towards improved retest reliability in High-Res OCT in iAMD eyes but without statistical significance. Improved axial resolution of the High-Res OCT benefits retest reliability of retinal layer annotation and improves perceived image quality and resolution. Automated image analysis algorithms could also benefit from the increased image resolution. Full article

(This article belongs to the Special Issue Biomedical Imaging and Analysis of the Eye)

► Show Figures

Graphical abstract

15 pages, 3880 KiB

Open AccessArticle

Detecting Early Ocular Choroidal Melanoma Using Ultrasound Localization Microscopy

by Biao Quan, Xiangdong Liu, Shuang Zhao, Xiang Chen, Xuan Zhang and Zeyu Chen

Bioengineering 2023, 10(4), 428; https://doi.org/10.3390/bioengineering10040428 - 28 Mar 2023

Cited by 2 | Viewed by 1682

Abstract

Ocular choroidal melanoma (OCM) is the most common ocular primary malignant tumor in adults, and there is an increasing emphasis on its early detection and treatment worldwide. The main obstacle in early detection of OCM is its overlapping clinical features with benign choroidal [...] Read more.

Ocular choroidal melanoma (OCM) is the most common ocular primary malignant tumor in adults, and there is an increasing emphasis on its early detection and treatment worldwide. The main obstacle in early detection of OCM is its overlapping clinical features with benign choroidal nevus. Thus, we propose ultrasound localization microscopy (ULM) based on the image deconvolution algorithm to assist the diagnosis of small OCM in early stages. Furthermore, we develop ultrasound (US) plane wave imaging based on three-frame difference algorithm to guide the placement of the probe on the field of view. A high-frequency Verasonics Vantage system and an L22-14v linear array transducer were used to perform experiments on both custom-made modules in vitro and a SD rat with ocular choroidal melanoma in vivo. The results demonstrate that our proposed deconvolution method implement more robust microbubble (MB) localization, reconstruction of microvasculature network in a finer grid and more precise flow velocity estimation. The excellent performance of US plane wave imaging was successfully validated on the flow phantom and in an in vivo OCM model. In the future, the super-resolution ULM, a critical complementary imaging modality, can provide doctors with conclusive suggestions for early diagnosis of OCM, which is significant for the treatment and prognosis of patients. Full article

(This article belongs to the Special Issue Biomedical Imaging and Analysis of the Eye)

► Show Figures

Graphical abstract

15 pages, 7870 KiB

Open AccessArticle

Rotational Distortion and Compensation in Optical Coherence Tomography with Anisotropic Pixel Resolution

by Guangying Ma, Taeyoon Son, Tobiloba Adejumo and Xincheng Yao

Bioengineering 2023, 10(3), 313; https://doi.org/10.3390/bioengineering10030313 - 01 Mar 2023

Cited by 2 | Viewed by 1324

Abstract

Accurate image registration is essential for eye movement compensation in optical coherence tomography (OCT) and OCT angiography (OCTA). The spatial resolution of an OCT instrument is typically anisotropic, i.e., has different resolutions in the lateral and axial dimensions. When OCT images have anisotropic [...] Read more.

Accurate image registration is essential for eye movement compensation in optical coherence tomography (OCT) and OCT angiography (OCTA). The spatial resolution of an OCT instrument is typically anisotropic, i.e., has different resolutions in the lateral and axial dimensions. When OCT images have anisotropic pixel resolution, residual distortion (

R D

) and false translation (

F T

) are always observed after image registration for rotational movement. In this study,

R D

and

F T

were quantitively analyzed over different degrees of rotational movement and various lateral and axial pixel resolution ratio (

R_{L} / R_{A}

) values. The

R D

and

F T

provide the evaluation criteria for image registration. The theoretical analysis confirmed that the

R D

and

F T

increase significantly with the rotation degree and

R_{L} / R_{A}

. An image resizing assisting registration (RAR) strategy was proposed for accurate image registration. The performance of direct registration (DR) and RAR for retinal OCT and OCTA images were quantitatively compared. Experimental results confirmed that unnormalized

R_{L} / R_{A}

causes

R D

and

F T

; RAR can effectively improve the performance of OCT and OCTA image registration and distortion compensation. Full article

(This article belongs to the Special Issue Biomedical Imaging and Analysis of the Eye)

► Show Figures

Graphical abstract

Review

Jump to: Research

24 pages, 14980 KiB

Open AccessReview

On Machine Learning in Clinical Interpretation of Retinal Diseases Using OCT Images

by Prakash Kumar Karn and Waleed H. Abdulla

Bioengineering 2023, 10(4), 407; https://doi.org/10.3390/bioengineering10040407 - 24 Mar 2023

Cited by 5 | Viewed by 2307

Abstract

Optical coherence tomography (OCT) is a noninvasive imaging technique that provides high-resolution cross-sectional retina images, enabling ophthalmologists to gather crucial information for diagnosing various retinal diseases. Despite its benefits, manual analysis of OCT images is time-consuming and heavily dependent on the personal experience [...] Read more.

Optical coherence tomography (OCT) is a noninvasive imaging technique that provides high-resolution cross-sectional retina images, enabling ophthalmologists to gather crucial information for diagnosing various retinal diseases. Despite its benefits, manual analysis of OCT images is time-consuming and heavily dependent on the personal experience of the analyst. This paper focuses on using machine learning to analyse OCT images in the clinical interpretation of retinal diseases. The complexity of understanding the biomarkers present in OCT images has been a challenge for many researchers, particularly those from nonclinical disciplines. This paper aims to provide an overview of the current state-of-the-art OCT image processing techniques, including image denoising and layer segmentation. It also highlights the potential of machine learning algorithms to automate the analysis of OCT images, reducing time consumption and improving diagnostic accuracy. Using machine learning in OCT image analysis can mitigate the limitations of manual analysis methods and provide a more reliable and objective approach to diagnosing retinal diseases. This paper will be of interest to ophthalmologists, researchers, and data scientists working in the field of retinal disease diagnosis and machine learning. By presenting the latest advancements in OCT image analysis using machine learning, this paper will contribute to the ongoing efforts to improve the diagnostic accuracy of retinal diseases. Full article

(This article belongs to the Special Issue Biomedical Imaging and Analysis of the Eye)

► Show Figures

Figure 1

Journal Menu

Journal Browser

Biomedical Imaging and Analysis of the Eye

Share This Special Issue

Special Issue Editors

Special Issue Information

Keywords

Related Special Issue

Published Papers (10 papers)

Research

Review

Further Information

Guidelines

MDPI Initiatives

Follow MDPI