RPE layer detection and baseline estimation using statistical methods and randomization for classification of AMD from retinal OCT

doi:10.1016/j.cmpb.2020.105822

Computer Methods and Programs in Biomedicine

Volume 200, March 2021, 105822

https://doi.org/10.1016/j.cmpb.2020.105822 Get rights and content

Highlights

•
Suitable adaptive denoising technique before estimation.
•
Modification of existing methodology to segment the RPE layer.
•
Novel method of estimating baseline for retinal OCT images using randomisation.
•
Completely automated with no requirements of seed points.
•
Works on individual OCT images and its performance is independent of volume size or data set size.
•
No requirement of training /annotated data.

Abstract

Background and Objective: Age-related macular degeneration (AMD) is a condition of the eye that affects the aged people. Optical coherence tomography (OCT) is a diagnostic tool capable of analyzing and identifying the disease affected retinal layers with high resolution. The objective of this work is to extract the retinal pigment epithelium (RPE) layer and the baseline (natural eye curvature, particular to every patient) from retinal spectral-domain OCT (SD-OCT) images. It uses them to find the height of drusen (abnormalities) in the RPE layer and classify it as AMD or normal.

Methods: In the proposed work, the contrast enhancement based adaptive denoising technique is used for speckle elimination. Pixel grouping and iterative elimination based on the knowledge of typical layer intensities and positions are used to obtain the RPE layer. Using this estimate, randomization techniques are employed, followed by polynomial fitting and drusen removal to arrive at a baseline estimate. The classification is based on the drusen height obtained by taking the difference between the RPE and baseline levels. We have used a patient, wise classification approach where a patient is classified diseased if more than a threshold number of patient images have drusen of more than a certain height. Since all slices of an affected patient will not show drusen, we are justified in adopting this technique.

Results: The proposed method is tested on a public data set of 2130 images/slices, which belonged to 30 patient volumes (15 AMD and 15 Normal) and achieved an overall accuracy of 96.66%, with no false positives. In comparison with existing works, the proposed method achieved higher overall accuracy and a better baseline estimate.

Conclusions: The proposed work focuses on AMD/normal classification using a statistical approach. It does not require any training. The proposed method modifies the motion restoration paradigm to obtain an application-specific denoising algorithm. The existing RPE detection algorithm is modified significantly to make it robust and applicable even to images where the RPE is not very evident/there is a significant amount of perforations (drusen). The baseline estimation algorithm employs a powerful combination of randomization, iterative polynomial fitting, and pixel elimination in contrast to mere fitting techniques. The main highlight of this work is, it achieved an exact estimation of the baseline in the retinal image compared to the existing methods.

Introduction

Optical coherence tomography (OCT) based imaging of the eye is widely accessible to a large population due to its low cost and zero side effects. The use of inexpensive and nonradioactive optical contrast agents can offer safe, highly sensitive, and targeted imaging of regions of interest. Portability, low instrumentation cost, and operational cost are other advantages of this mechanism [1]. There are many works carried out in eye disease detection using OCT images [2], [3], [4], [5], [6], [7]. 8.7% of the worldwide population has Age-related macular degeneration (AMD), and the projected number of people with the disease is around 196 million in 2020, increasing to 288 million in 2040 [8]. AMD can be dry and wet forms. The term ”dry AMD” refers broadly to early or intermediate stages, and a late-stage referred to as geographic atrophy (GA). The advanced GA stage involves the loss of RPE and choroid in at least the retina’s macular region, which leads to a gradual loss of photoreceptors and central vision [9]. There is an abnormal growth of blood vessels under the macula in the wet form, referred to as choroidal neovascularization, which pours out blood and fluid into the retina. Statistics show that only 10% of AMD patients suffer from wet form [10].

Patients suffering from dry AMD have a significant anomaly in the Retinal pigment epithelium (RPE) layer. This Macular retinal anomaly is known as drusen [11]. The drusen appear as a bulge/perforation. The greater the severity of the problem, the more significant is the drusen. It can be visualized with OCT. Literature suggests several approaches to simultaneous segmentation of all layers, by using data from the A-scans in a particular patient volume [12]. Though the process is applicable even to low contrast OCT images, it is only semi-automated. It is also not very effective when patient volumes are small, and there is a large variance between the patient’s A-scans. Denoising images are challenging as the noise intensity levels, and the extent of prevalence vary from machine to machine. The denoising algorithm should be specific, depending on the subsequent step/layer extracted to give better results. The proposed work comprises a denoising technique that helps in RPE extraction, followed by RPE detection and baseline estimation for AMD classification.

The main features of the work are:

1.
Suitable contrast enhancement based adaptive denoising technique before estimation
2.
Modification of existing methodology to segment the RPE layer
3.
Novel method of estimating baseline for retinal OCT images using randomization
4.
Completely automated with no requirements of seed points
5.
Works on individual OCT images, performance scales linearly with image size, and does not depend on a larger set of labelled training data or number of patient images in a volume.

Very few slices of a patient who is suffering from AMD show drusen. Hence for training, a network, handpicking the data is tedious. This mechanism can also help automate such a process by identifying RPE perforated images.

Section snippets

Related works

Several papers deal with the segmentation of the RPE layer in particular for AMD classification. The authors of [13] proposed a stepwise algorithm to obtain the RPE and Retinal Nerve Fiber layer (RNFL), following which drusen and bubble detection were done to extract relevant features from the image for classifications. The three stages of classification were Diabetic Macular Edema (DME), AMD, and unaffected. However, testing was done on only 16 OCT images. A similar stepwise approach was used

Methods

As discussed earlier, OCT is a powerful, non-invasive, and affordable imaging technique that can be used to detect diseases like AMD and glaucoma. A four-step algorithm encompassing despeckling, RPE detection, baseline estimation, and classification of a given patient’s OCT image is employed. The methodology adopted is shown in Fig. 1.

Results

The proposed method is tested on Duke’s publicly available data set [17]. The data set consists of 15 diseased and 15 healthy patients. There are a total of 2130 slices. The number of slices per patient ranged from 37 to 97. The proposed work achieves an overall accuracy of 96.66% by considering all of the data set images. The confusion matrix for the proposed work is shown in Table 1. Precision, Recall, and F1- Score is common classification metrics which have been calculated and presented in

Funding

This research work is funded by Science and Engineering Research Board (SERB), Government of India, Grant No. SERB/F/11639/2018-2019 dated 26 February 2019.

Declaration of Competing Interest

We wish to draw the attention of the Editor to the following facts which may be considered as potential conflicts of interest and to significant financial contributions to this work. We confirm that the manuscript has been read and approved by all named authors and that there are no other persons who satisfied the criteria for authorship but are not listed. We further confirm that the order of authors listed in the manuscript has been approved by all of us.

We confirm that we have given due

Acknowledgment

The authors would like to thank Dr. S Sujatha of Institute of Ophthalmology, Joseph Eye Hospital, Tiruchirappalli, India for her valuable guidance throughout the research.

References (25)

P. Sudeep et al.
Enhancement and bias removal of optical coherence tomography images: an iterative approach with adaptive bilateral filtering
Comput. Biol. Med.
(2016)
S.I. Niwas et al.
Automated anterior segment oct image analysis for angle closure glaucoma mechanisms classification
Comput Methods Programs Biomed
(2016)
S.I. Niwas et al.
Complex wavelet based quality assessment for as-oct images with application to angle closure glaucoma diagnosis
Comput Methods Programs Biomed
(2016)
W.L. Wong et al.
Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis
The Lancet Global Health
(2014)
Z. Yaqoob et al.
Methods and application areas of endoscopic optical coherence tomography
J Biomed Opt
(2006)
K. Jemshi et al.
Development of an efficient algorithm for the detection of macular edema from optical coherence tomography images
Int J Comput Assist Radiol Surg
(2018)
S.I. Niwas et al.
Cross-examination for angle-closure glaucoma feature detection
IEEE J Biomed Health Inform
(2015)
S.I. Niwas et al.
Reliable feature selection for automated angle closure glaucoma mechanism detection
J Med Syst
(2015)
C.A. Curcio et al.
Photoreceptor loss in age-related macular degeneration.
Investigative ophthalmology & visual science
(1996)
M.D. Davis et al.
The age-related eye disease study severity scale for age-related macular degeneration: areds report no. 17.
Archives of ophthalmology (Chicago, Ill.: 1960)
(2005)

S. Naz et al.

Automated segmentation of rpe layer for the detection of age macular degeneration using oct images

2016 Sixth International Conference on Image Processing Theory, Tools and Applications (IPTA)

(2016)

W. Duan et al.

A generative model for oct retinal layer segmentation by groupwise curve alignment

IEEE Access

(2018)

Cited by (20)

Towards more efficient ophthalmic disease classification and lesion location via convolution transformer
2022, Computer Methods and Programs in Biomedicine
Citation Excerpt :
With machine learning approaches, Thomas et al. proposed a method to detect the retinal pigment epithelium (RPE) layer and use statistical methods and randomization to analyze AMD diseases. The method has relatively high accuracy in detecting drusen and requires no training, but some thresholds must be set manually [12]. Srinivasan et al. used the multi-scale histogram of the oriented gradient descriptor (HOG) to extract features from DME, dry AMD, and normal images.
A retina optical coherence tomography (OCT) image differs from a traditional image due to its significant speckle noise, irregularity, and inconspicuous features. A conventional deep learning architecture cannot effectively improve the classification accuracy, sensitivity, and specificity of OCT images, and noisy images are not conducive to further diagnosis. This paper proposes a novel lesion-localization convolution transformer (LLCT) method, which combines both convolution and self-attention to classify ophthalmic diseases more accurately and localize the lesions in retina OCT images.
A novel architecture design is accomplished through applying customized feature maps generated by convolutional neutral network (CNN) as the input sequence of self-attention network. This design takes advantages of CNN's extracting image features and transformer's consideration of global context and dynamic attention. Part of the model is backward propagated to calculate the gradient as a weight parameter, which is multiplied and summed with the global features generated by the forward propagation process to locate the lesion.
Extensive experiments show that our proposed design achieves improvement of about 7.6% in overall accuracy, 10.9% in overall sensitivity, and 9.2% in overall specificity compared with previous methods. And the lesions can be localized without the labeling data of lesion location in OCT images.
The results prove that our method significantly improves the performance and reduces the computation complexity in artificial intelligence assisted analysis of ophthalmic disease through OCT images.
Our method has a significance boost in ophthalmic disease classification and location via convolution transformer. This is applicable to assist ophthalmologists greatly.¹
Redundancy reduced depthwise separable convolution for glaucoma classification using OCT images
2022, Biomedical Signal Processing and Control
Citation Excerpt :
Compared to TD-OCT, SD-OCT is more sensitive in detecting RNFL deformations in the progression of glaucoma [26]. OCT provides us with a clear view of the nine retinal layers, as shown in Fig. 1, whose varying thickness and irregularities encode a lot of information that is helpful in the diagnosis of diseases, such as Choroidal Neovascularization (CNV) Diabetic Macular Edema (DME) and drusen and glaucoma [27–34]. Fig. 2 represents the retinal layer deformation in the OCT image of an individual with glaucoma in general.
Glaucoma is a chronic progressive optic neuropathy characterized by the impairment of the optic nerve, and if it is left untreated, it may lead to irreversible vision loss. So, the early detection, diagnosis, and management of retinal glaucoma are necessary as the number of cases increases expeditiously. Existing deep learning classification methods of glaucoma are computationally intensive, restricting memory potency, training, and affects the optimization of hyperparameters. Thus they are unsuitable for real-time applications with limited computing resources. A two-dimensional depthwise separable convolution architecture can significantly improve the efficiency of parameter utilization and calculation speed. This paper proposes a raw SD-OCT-based depthwise separable convolution model to classify glaucoma from healthy images. Each input channel is convolved with each filter kernel, and the resulting output channels are effectively mixed by pointwise convolution. A gradient-weighted class activation mapping is estimated to highlight the region with structural deformations per B scan to validate the performance qualitatively. Our private database comprises 1105 glaucomatous and 1049 normal OCT B scans around ONH, which aids the ophthalmologists in making an appropriate diagnosis. The proposed redundancy reduced depthwise separable convolution network achieved an accuracy, precision, recall, F1-score, and AUC of 0.9963, 0.9946, 0.9982, 0.9964, and 0.9963, respectively. The number of parameters used in the proposed method is only 20,686. In the proposed network, tedious segmentation and other image processing steps are avoided, still achieving remarkable network training efficacy with appreciable reduction of learnable parameters.
Deep learning based joint segmentation and characterization of multi-class retinal fluid lesions on OCT scans for clinical use in anti-VEGF therapy
2021, Computers in Biology and Medicine
Citation Excerpt :
Since the last two decades, the field of retinal image processing has received much attention, with an emphasis on OCT-based automated retinal analysis. Researchers have suggested several automated algorithms for the following main tasks relevant to retinal OCT imaging: (1) elimination of speckles and OCT denoising [17,18], (2) chorioretinal layers segmentation [19–21], (3) segmentation of anatomical lesions [21–25], and (4) classification of retinal pathologies [25–30]. The segmentation of MRF lesions in OCT imaging is realized as a difficult task for two main reasons: 1) high signal-to-noise ratio (SNR) variability between the scans of different OCT vendors (Topcon, Cirrus, and Spectralis); and 2) the considerable discrepancies in volume, structure, and delineation of the MRF lesions [13].
In anti-vascular endothelial growth factor (anti-VEGF) therapy, an accurate estimation of multi-class retinal fluid (MRF) is required for the activity prescription and intravitreal dose. This study proposes an end-to-end deep learning-based retinal fluids segmentation network (RFS-Net) to segment and recognize three MRF lesion manifestations, namely, intraretinal fluid (IRF), subretinal fluid (SRF), and pigment epithelial detachment (PED), from multi-vendor optical coherence tomography (OCT) imagery. The proposed image analysis tool will optimize anti-VEGF therapy and contribute to reducing the inter- and intra-observer variability.
The proposed RFS-Net architecture integrates the atrous spatial pyramid pooling (ASPP), residual, and inception modules in the encoder path to learn better features and conserve more global information for precise segmentation and characterization of MRF lesions. The RFS-Net model is trained and validated using OCT scans from multiple vendors (Topcon, Cirrus, Spectralis), collected from three publicly available datasets. The first dataset consisted of OCT volumes obtained from 112 subjects (a total of 11,334 B-scans) is used for both training and evaluation purposes. Moreover, the remaining two datasets are only used for evaluation purposes to check the trained RFS-Net's generalizability on unseen OCT scans. The two evaluation datasets contain a total of 1572 OCT B-scans from 1255 subjects. The performance of the proposed RFS-Net model is assessed through various evaluation metrics.
The proposed RFS-Net model achieved the mean F₁ scores of 0.762, 0.796, and 0.805 for segmenting IRF, SRF, and PED. Moreover, with the automated segmentation of the three retinal manifestations, the RFS-Net brings a considerable gain in efficiency compared to the tedious and demanding manual segmentation procedure of the MRF.
Our proposed RFS-Net is a potential diagnostic tool for the automatic segmentation of MRF (IRF, SRF, and PED) lesions. It is expected to strengthen the inter-observer agreement, and standardization of dosimetry is envisaged as a result.
A novel multiscale and multipath convolutional neural network based age-related macular degeneration detection using OCT images
2021, Computer Methods and Programs in Biomedicine
Background and Objective: One of the significant retinal diseases that affected older people is called Age-related Macular Degeneration (AMD). The first stage creates a blur effect on vision and later leads to central vision loss. Most people overlooked the primary stage blurring and converted it into an advanced stage. There is no proper treatment to cure the disease. So the early detection of AMD is essential to prevent its extension into the advanced stage. This paper proposes a novel deep Convolutional Neural Network (CNN) architecture to automate AMD diagnosis early from Optical Coherence Tomographic (OCT) images.
Methods: The proposed architecture is a multiscale and multipath CNN with six convolutional layers. The multiscale convolution layer permits the network to produce many local structures with various filter dimensions. The multipath feature extraction permits CNN to merge more features regarding the sparse local and fine global structures. The performance of the proposed architecture is evaluated through ten-fold cross-validation methods using different classifiers like support vector machine, multi-layer perceptron, and random forest.
Results: The proposed CNN with the random forest classifier gives the best classification accuracy results. The proposed method is tested on data set 1, data set 2, data set 3, data set 4, and achieved an accuracy of 0.9666, 0.9897, 0.9974, and 0.9978 respectively, with random forest classifier. Also, we tested the combination of first three data sets and achieved an accuracy of 0.9902.
Conclusions: An efficient algorithm for detecting AMD from OCT images is proposed based on a multiscale and multipath CNN architecture. Comparison with other approaches produced results that exhibit the efficiency of the proposed algorithm in the detection of AMD. The proposed architecture can be applied in rapid screening of the eye for the early detection of AMD. Due to less complexity and fewer learnable parameters.
A novel multiscale convolutional neural network based age-related macular degeneration detection using OCT images
2021, Biomedical Signal Processing and Control
Citation Excerpt :
Using retinal OCT images for diagnosis purposes is advantageous over fundus images. OCT provides us with a clear view of the nine retinal layers, as shown in Fig. 1, whose varying thickness and irregularities encode a lot of information that is helpful in the diagnosis of diseases, such as choroidal neovascularization (CNV) diabetic macular edema (DME) and drusen [4–8]. Besides ophthalmic diagnosis, recent studies have shown OCT images contain bio-markers for many other diseases such as Parkinson's disease [9], Alzheimer's disease [10], etc.
Age-related macular degeneration (AMD) is an ocular disorder that affects the elderly. The prevalence of AMD is growing due to the aging population in society; hence early diagnosis is necessary to prevent vision loss in the elderly. Arranging a detailed eye screening system for detecting AMD is a very challenging process. This paper proposes a novel multiscale convolutional neural network (CNN) architecture for accurate diagnosis of AMD. The architecture proposed is a multiscale CNN with seven convolutional layers for classifying AMD or normal images. The multiscale convolution layer enables a large number of local structures to be generated with various filter sizes. In this proposed network, the sigmoid function is used as the classifier. The proposed CNN network is trained on the Mendeley dataset and tested on four datasets, namely Mendeley, OCTID, Duke, SD-OCT Noor dataset and achieved an accuracy of 99.73%, 98.08%, 96.66%, and 97.95% respectively. Comparison with alternative methods yielded results that exhibit the efficiency of the proposed algorithm in AMD detection. Even if the proposed model is trained only on the Mendeley dataset, it achieved good detection accuracy when tested with other datasets. This indicates the proposed model's ability to classify AMD/Normal images from other datasets. Comparison with other approaches produced results that exhibit the efficiency of the proposed algorithm in detecting AMD. The proposed architecture can be applied in rapid screening of the eye for the early detection of AMD. Due to less complexity and fewer learnable parameters, the proposed CNN can be implemented in real-time.
Wavelet scattering transform application in classification of retinal abnormalities using OCT images
2023, Scientific Reports

View all citing articles on Scopus

View full text

RPE layer detection and baseline estimation using statistical methods and randomization for classification of AMD from retinal OCT

Highlights

Abstract

Introduction

Section snippets

Related works

Methods

Results

Funding

Declaration of Competing Interest

Acknowledgment

Comput. Biol. Med.

Comput Methods Programs Biomed

Comput Methods Programs Biomed

The Lancet Global Health

Methods and application areas of endoscopic optical coherence tomography

J Biomed Opt

Development of an efficient algorithm for the detection of macular edema from optical coherence tomography images

Int J Comput Assist Radiol Surg

Cross-examination for angle-closure glaucoma feature detection

IEEE J Biomed Health Inform

Reliable feature selection for automated angle closure glaucoma mechanism detection

J Med Syst

Photoreceptor loss in age-related macular degeneration.

Investigative ophthalmology & visual science

The age-related eye disease study severity scale for age-related macular degeneration: areds report no. 17.

Archives of ophthalmology (Chicago, Ill.: 1960)

Automated segmentation of rpe layer for the detection of age macular degeneration using oct images

2016 Sixth International Conference on Image Processing Theory, Tools and Applications (IPTA)

A generative model for oct retinal layer segmentation by groupwise curve alignment

IEEE Access