SEMS-DRNet: Attention enhanced multi-scale residual blocks with Bayesian optimization for diabetic retinopathy classification

Bibliometric analysis is a rigorous method to analyze significant quantities of bibliometric data to assess their impact on a particular field. This study used bibliometric analysis to investigate the academic research on diabetes detection and classification from 2000 to 2023. The PRISMA 2020 framework was followed to identify, filter, and select relevant papers. This study used the Web of Science database to determine relevant publications concerning diabetes detection and classification using the keywords "diabetes detection", "diabetes classification", and "diabetes detection and classification". A total of 863 publications were selected for analysis. The research applied two bibliometric techniques: performance analysis and science mapping. Various bibliometric parameters, including publication analysis, trend analysis, citation analysis, and networking analysis, were used to assess the performance of these articles. The analysis findings showed that India, China, and the United States are the top three countries with the highest number of publications and citations on diabetes detection and classification. The most frequently used keywords are machine learning, diabetic retinopathy, and deep learning. Additionally, the study identified "classification", "diagnosis", and "validation" as the prevailing topics for diabetes identification. This research contributes valuable insights into the academic landscape of diabetes detection and classification.

Diabetes is a globally prevalent disease that can cause microvascular compilation such as diabetic retinopathy (DR) in the human eye organs, and it might be a significant reason for visual deficiency. The present study aims to develop an automated diabetic retinopathy detection and classification system from digital fundus images to reduce the workload of ophthalmologists. This work comprises three main stages. Our method starts with the extraction of blood vessels from retinal images using Kirsch’s templates. Secondly, the method detects whatever the input images as normal or diabetic retinopathy and then illustrates an automatic diabetic retinopathy classification technique through statistical texture features. It embeds gray-level co-occurrence matrix (GLCM) and gray-level run length matrix (GLRLM) for second-order and higher-order statistical texture features as a feature extraction technique into three-renowned classifiers namely K-nearest neighbor (KNN), random forest (RF), and support vector machine (SVM). The evaluation results containing a dataset of 644 retinal images indicate that the proposed method based on random forest classifier is found to be effective with a weighted sensitivity, precision, F1-score, and accuracy of 95.53%, 96.45%, 95.38%, and 95.19%, respectively. It is expected that the suggested medical decision support system will be able to detect and classify diabetic retinopathy well.

Because of retina abnormalities of diabetic patients, the most common vision-threatening disease is diabetic retinopathy (DR). The DR diagnosis and prevention are challenging tasks as they may lead to vision loss. According to the literature analysis, the shortcomings in existing studies, such as failed to reduce the feature dimension, higher execution time, and higher computational cost, unable to tune the hyper-parameters, such as a number of hidden layers and learning rate, more computational complexities, higher cost, and so forth, during DR classification. To tackle these problems, we proposed a deep long- and short-term memory (LSTM) in a neural network with Red Fox optimization (deep LSTM-RFO) algorithm for DR classification. The four major components involved in the proposed methods are image preprocessing, segmentation, feature extraction, and classification. At first, an adaptive histogram equalization and histogram equalization model performs the fundus image preprocessing, thereby neglecting the noise and improving the contrast level of an image. Next, an adaptive watershed segmentation model effectively segments the lesion region based on the optic disc color and size of hemorrhages. At the third stage, we have extracted statistical, intensity, color, and shape features. Finally, the single normal class with three abnormal classes such as mild non-proliferative diabetic retinopathy, moderate NPDR, and severe NPDR are accurately classified using the deep LSTM-RFO algorithm. Experimentally, the MESSIDOR, STARE, and DRIVE datasets are used for both training and validation. MATLAB software performs the implementation process with respect to various evaluation criteria used. However, the proposed method accomplished superior performance, such as 98.45% specificity, 96.78% sensitivity, 97.92% precision, 96.89% recall, and 97.93% F-score results in terms of DR classification than previous methods.

Background: Despite advances in artificial intelligence (AI) for Diabetic Retinopathy (DR) classification, traditional severity-based approaches often lack interpretability and fail to capture specific lesion-centered characteristics. To address these limitations, we constructed the National Medical Center (NMC) dataset, independently annotated by medical professionals with detailed labels of major DR lesions, including retinal hemorrhages, microaneurysms, and exudates. Methods: This study explores four critical research questions. First, we assess the analytical advantages of lesion-centered labeling compared to traditional severity-based labeling. Second, we investigate the potential complementarity between these labeling approaches through integration experiments. Third, we analyze how various model architectures and classification strategies perform under different labeling schemes. Finally, we evaluate decision-making differences between labeling methods using visualization techniques. We benchmarked the lesion-centered NMC dataset against the severity-based public Asia Pacific Tele-Ophthalmology Society (APTOS) dataset, conducting experiments with EfficientNet—a convolutional neural network architecture—and diverse classification strategies. Results: Our results demonstrate that binary classification effectively identifies severe non-proliferative Diabetic Retinopathy (Severe NPDR) exhibiting complex lesion patterns, while relationship-based learning enhances performance for underrepresented classes. Transfer learning from NMC to APTOS notably improved severity classification, achieving performance gains of 15.2% in mild cases and 66.3% in severe cases through feature fusion using Bidirectional Feature Pyramid Network (BiFPN) and Feature Pyramid Network (FPN). Visualization results confirmed that lesion-centered models focus more precisely on pathological features. Conclusions: Our findings highlight the benefits of integrating lesion-centered and severity-based information to enhance both accuracy and interpretability in DR classification. Future research directions include spatial lesion mapping and the development of clinically grounded learning methodologies.

Domain Generalization is a challenging problem in deep learning especially in medical image analysis because of the huge diversity between different datasets. Existing papers in the literature tend to optimize performance on single target domains, without regards to model generalizability on other domains or distributions. High discrepancy in the number of images and major domain shifts, can therefore cause single-source trained models to under-perform during testing. In this paper, we address the problem of domain generalization in Diabetic Retinopathy (DR) classification. The baseline for comparison is set as joint training on different datasets, followed by testing on each dataset individually. We therefore introduce a method that encourages seeking a flatter minima during training while imposing a regularization. This reduces gradient variance from different domains and therefore yields satisfactory results on out-of-domain DR classification. We show that adopting DR-appropriate augmentations enhances model performance and in-domain generalizability. By performing our evaluation on 4 open-source DR datasets, we show that the proposed domain generalization method outperforms separate and joint training strategies as well as well-established methods. Source Code is available at https://github.com/BioMedIA-MBZUAI/DRGen.KeywordsDeep learningDiabetic retinopathyDomain generalizationRegularization

Diabetic Retinopathy (DR) has become a widespread illness among diabetics across the globe. Retinal fundus images are generally used by physicians to detect and classify the stages of DR. Since manual examination of DR images is a time-consuming process with the risks of biased results, automated tools using Artificial Intelligence (AI) to diagnose the disease have become essential. In this view, the current study develops an Optimal Deep Learning-enabled Fusion-based Diabetic Retinopathy Detection and Classification (ODL-FDRDC) technique. The intention of the proposed ODL-FDRDC technique is to identify DR and categorize its different grades using retinal fundus images. In addition, ODL-FDRDC technique involves region growing segmentation technique to determine the infected regions. Moreover, the fusion of two DL models namely, CapsNet and MobileNet is used for feature extraction. Further, the hyperparameter tuning of these models is also performed via Coyote Optimization Algorithm (COA). Gated Recurrent Unit (GRU) is also utilized to identify DR. The experimental results of the analysis, accomplished by ODL-FDRDC technique against benchmark DR dataset, established the supremacy of the technique over existing methodologies under different measures.

Recently, Telehealth connects patients to vital healthcare services via remote monitoring, wireless communications, videoconferencing, and electronic consults. By increasing access to specialists and physicians, telehealth assists in ensuring patients receive the proper care at the right time and right place. Teleophthalmology is a study of telemedicine that provides services for eye care using digital medical equipment and telecommunication technologies. Multimedia computing with Explainable Artificial Intelligence (XAI) for telehealth has the potential to revolutionize various aspects of our society, but several technical challenges should be resolved before this potential can be realized. Advances in artificial intelligence methods and tools reduce waste and wait times, provide service efficiency and better insights, and increase speed, the level of accuracy, and productivity in medicine and telehealth. Therefore, this study develops an XAI-enabled teleophthalmology for diabetic retinopathy grading and classification (XAITO-DRGC) model. The proposed XAITO-DRGC model utilizes OphthoAI IoMT headsets to enable remote monitoring of diabetic retinopathy (DR) disease. To accomplish this, the XAITO-DRGC model applies median filtering (MF) and contrast enhancement as a pre-processing step. In addition, the XAITO-DRGC model applies U-Net-based image segmentation and SqueezeNet-based feature extractor. Moreover, Archimedes optimization algorithm (AOA) with a bidirectional gated recurrent convolutional unit (BGRCU) is exploited for DR detection and classification. The experimental validation of the XAITO-DRGC method can be tested using a benchmark dataset and the outcomes are assessed under distinct prospects. Extensive comparison studies stated the enhancements of the XAITO-DRGC model over recent approaches.

Automated diagnosis of diabetic retinopathy from fundus images involves detecting both small- and large-scale lesions, which makes this a difficult task for deep learning applications. In this paper we investigate the effects of small scale feature propagation for improving diabetic retinopathy classification. To accomplish this, we have utilized a publicly available dataset with 88,702 images, which contains unbalanced number of examples for different classes. A linear equation for class-specific gradient weighting has been proposed and has found to be beneficial. Three different residual architectures with residual and skip connections have been tested and their efficacy for this task is examined. The residual connections have been verified to improve the results for detecting small scale features for deep architectures. Skip connections within the current experimental setting have been found to be detrimental for the overall performance, potential solutions and their resulting effects have been discussed.

Diabetes is characterized by impaired metabolism of glucose caused by insulin deficiency. Diabetic retinopathy is the eye disease, is caused by retinal damage which is generally formed as a result of diabetes mellitus. It is a serious vascular disorder for which early detection and the treatment are required to inhibit the intense vision loss. Also, the diagnosis entails skilled professionals for detection because non-automatic screening methods are very time consuming and are not that efficient for a large number of retinal images. This paper provides a broad review of various techniques and methodologies used by the authors for diabetic retinopathy detection and classification. Furthermore, most recent work and developments are studied in this paper. We are proposing an advanced deep learning CNN approach for automatic diagnosis of DR from color fundus images.

Automated Diabetic Retinopathy detection and classification using stochastic coordinate descent deep learning architectures

Chapter 3 - Classification of Diabetic Retinopathy

AimsThis study aims to compare the performance of a handheld fundus camera (Eyer) and standard tabletop fundus cameras (Visucam 500, Visucam 540, and Canon CR-2) for diabetic retinopathy and diabetic macular edema screening.MethodsThis was a multicenter, cross-sectional study that included images from 327 individuals with diabetes. The participants underwent pharmacological mydriasis and fundus photography in two fields (macula and optic disk centered) with both strategies. All images were acquired by trained healthcare professionals, de-identified, and graded independently by two masked ophthalmologists, with a third senior ophthalmologist adjudicating in discordant cases. The International Classification of Diabetic Retinopathy was used for grading, and demographic data, diabetic retinopathy classification, artifacts, and image quality were compared between devices. The tabletop senior ophthalmologist adjudication label was used as the ground truth for comparative analysis. A univariate and stepwise multivariate logistic regression was performed to determine the relationship of each independent factor in referable diabetic retinopathy.ResultsThe mean age of participants was 57.03 years (SD 16.82, 9–90 years), and the mean duration of diabetes was 16.35 years (SD 9.69, 1–60 years). Age (P = .005), diabetes duration (P = .004), body mass index (P = .005), and hypertension (P < .001) were statistically different between referable and non-referable patients. Multivariate logistic regression analysis revealed a positive association between male sex (OR 1.687) and hypertension (OR 3.603) with referable diabetic retinopathy. The agreement between devices for diabetic retinopathy classification was 73.18%, with a weighted kappa of 0.808 (almost perfect). The agreement for macular edema was 88.48%, with a kappa of 0.809 (almost perfect). For referable diabetic retinopathy, the agreement was 85.88%, with a kappa of 0.716 (substantial), sensitivity of 0.906, and specificity of 0.808. As for image quality, 84.02% of tabletop fundus camera images were gradable and 85.31% of the Eyer images were gradable.ConclusionsOur study shows that the handheld retinal camera Eyer performed comparably to standard tabletop fundus cameras for diabetic retinopathy and macular edema screening. The high agreement with tabletop devices, portability, and low costs makes the handheld retinal camera a promising tool for increasing coverage of diabetic retinopathy screening programs, particularly in low-income countries. Early diagnosis and treatment have the potential to prevent avoidable blindness, and the present validation study brings evidence that supports its contribution to diabetic retinopathy early diagnosis and treatment.

Purpose Diabetic retinopathy is an ophthalmic disease that impairs the retinal blood vessels. Diabetic retinopathy can lead to blindness when it is not examined in earlier phases. Adversely, the accurate diabetic retinopathy recognition phase is prominently complicated and needs experienced human analysis of fundus images. Blockchain technology helps share data by allowing users to select what information to share and control who can access it, which is important for managing electronic health records in healthcare sector. Nevertheless, the privacy of user data is compromised due to the training data, which is revealed to unauthorized users. Methods In this work, a superior module for diabetic retinopathy classification based on Blockchain using principal convolutional analysis neural network is designed. Here, the simulation of Blockchain is carried out. Here, the input image is pre-processed using the Gaussian filter. LadderNet is deployed for lesion segmentation, and the segmentation of blood vessel is done using the Sine-Net model. Moreover, feature extraction is performed with the input image, lesion-segmented image, and blood vessel-segmented image. Finally, diabetic retinopathy classification is executed utilizing the proposed principal convolutional analysis neural network, which is classified into normal, mild non-proliferative diabetic retinopathy, moderate non-proliferative diabetic retinopathy, severe non-proliferative diabetic retinopathy, and proliferative. Results The Blockchain enabled principal convolutional analysis neural network obtains superior values of 90.9%, 91.9%, 92.5%, 89.4%, 88.4%, and 75.5% in terms of metrics like accuracy, true positive rate, true negative rate, positive predictive value, negative predictive value, and Mathews correlation coefficient. Conclusion The integration of principal convolutional analysis neural network with Blockchain enhances data integrity and patient privacy, making it a promising solution for early diagnosis and treatment. Also, this approach ensures accurate and efficient detection of diabetic retinopathy.

Diabetic retinopathy is a vision-threatening complication of diabetes that affects the retina, the light-sensitive tissue at the back of the eye. This condition arises as a result of prolonged high blood sugar levels, which can damage the small blood vessels in the retina. Diabetic retinopathy typically progresses through different stages, starting with mild non-proliferative retinopathy, where small blood vessels in the retina become weakened and leak. The classification of diabetic retinopathy plays a fundamental role in assessing the effectiveness of treatment and monitoring the progression of the disease over time, ultimately contributing to the preservation of patients’ vision and their overall quality of life. This research paper presents efficient technique for diabetic retinopathy (DR) classification and grading using data augmentation and a dynamic weighted optimization approach. The study contributes to the field of DR in several significant ways. Firstly, advanced data augmentation techniques are employed to generate diverse and representative features from retinal fundus images, enhancing the robustness and generalization capabilities of the models. Secondly, novel segmentation approaches, including multi-level Otsu thresholding and morphological operations, accurately localize and isolate affected regions in retinal images. Thirdly, innovative feature extraction and selection methods, such as Gray-Level Co-occurrence Matrix (GLCM) and dynamic Flemingo optimization, improve the selection of discriminative features for DR classification. Additionally, a novel cascaded voting ensemble deep neural network model is introduced, which combines the predictions of multiple learning algorithms to enhance classification performance. Lastly, the research addresses the grading of diabetic retinopathy by aligning the classification results with a standardized grading system, providing clinicians with accurate severity assessments for effective treatment decisions. Overall, this papers offers valuable insights and methodologies for improving the classification and grading of diabetic retinopathy, thereby contributing to the advancement of diagnosis and management strategies in the field.

To apply machine learning (ML) algorithms to perform multiclass diabetic retinopathy (DR) classification using both clinical data and optical coherence tomography angiography (OCTA). In this cross-sectional observational study, clinical data and OCTA parameters from 203 diabetic patients (203 eye) were used to establish the ML models, and those from 169 diabetic patients (169 eye) were used for independent external validation. The random forest, gradient boosting machine (GBM), deep learning and logistic regression algorithms were used to identify the presence of DR, referable DR (RDR) and vision-threatening DR (VTDR). Four different variable patterns based on clinical data and OCTA variables were examined. The algorithms' performance were evaluated using receiver operating characteristic curves and the area under the curve (AUC) was used to assess predictive accuracy. The random forest algorithm on OCTA+clinical data-based variables and OCTA+non-laboratory factor-based variables provided the higher AUC values for DR, RDR and VTDR. The GBM algorithm produced similar results, albeit with slightly lower AUC values. Leading predictors of DR status included vessel density, retinal thickness and GCC thickness, as well as the body mass index, waist-to-hip ratio and glucose-lowering treatment. ML-based multiclass DR classification using OCTA and clinical data can provide reliable assistance for screening, referral, and management DR populations.