Transfer Learning with Convolutional Neural Networks for Diabetic Retinopathy Image Classification. A Review

Diabetic retinopathy (DR) is a serious complication of diabetes that can result in vision loss if untreated, often progressing silently without warning symptoms. Elevated blood glucose levels damage the retina's microvasculature, initiating the condition. Early detection through retinal fundus imaging, supported by timely analysis and treatment, is critical for managing DR effectively. However, manually inspecting these images is a labour-intensive and time-consuming process, making computer-aided diagnosis (CAD) systems invaluable in supporting ophthalmologists. This research introduces the Fundus Imaging Diabetic Retinopathy Classification using Deep Learning and Fennec Fox Optimization (FIDRC-DLFFO) model, which automates the identification and classification of DR. The model integrates several advanced techniques to enhance performance and accuracy.1.The proposed FIDRC-DLFFO model automates DR detection and classification by combining median filtering for noise reduction, Inception-ResNet-v2 for feature extraction, and a gated recurrent unit (GRU) for classification.2.Fennec Fox Optimization (FFO) fine-tunes the GRU hyperparameters, boosting classification accuracy, with its effectiveness demonstrated on benchmark datasets.3.The results provide insights into the model's effectiveness and potential for real-world application.

Diabetic retinopathy (DR) is a complication of diabetes that cause retinal damage; therefore, it is a leading cause of blindness. However, early detection of this disease can dramatically reduce the risk of vision loss. The main problem of early DR detection is that the manual diagnosis by ophthalmology is time-consuming, expensive, and prone to misdiagnosis. Deep learning (DL) models have aided in the early diagnosis of DR, and DL is now frequently utilized in DR detection and classification. The main issues with classical DL models is that they are incapable to quantify the uncertainty in the models, thus they are prone to make wrong decisions in complex cases. However, Bayesian deep learning (BDL) models have recently evolved as unified probabilistic framework to integrate DL and Bayesian models to provides an accurate framework to identify all sources of uncertainty in the model. This paper introduces BDL and most recent research that used BDL approaches to treat diabetic retinopathy are reviewed and discussed. A thorough comparison of the existing Bayesian approaches in this topic is also presented. In addition, available datasets for the fundus retina, which is often employed in DR, are provided and reviewed.

Early detection of Diabetic Retinopathy (DR) is vital for preserving vision and preventing deterioration of eyesight. Fundus Fluorescein Angiography (FFA), recognized as the gold standard for diagnosing DR, effectively reveals abnormalities in retinal vasculature. Given the labor-intensive and costly nature of manual DR diagnosis, along with its low accuracy, developing a DR classification model based on FFA using deep learning techniques is crucial. Furthermore, DR classification faces challenges such as minimal lesion variance between different disease stages and significant size variations of lesions within the same stage, with small lesions often overlooked by existing models. We propose a deep learning model, D-GET, utilizing a Group-Enhanced Transformer for classifying DR lesion severity in FFA images. The D-GET model incorporates a Full-Scale Transformer Block, where the Group-Focal module captures feature information at multiple scales, from fine details to broader patterns, and adaptively integrates contextual information, enhancing the model's ability to detect small-scale lesions. The model also includes a Channel Adaptive Attention Module (CAAM) that synthesizes channel and spatial information to improve feature detection and localization. Experimental findings indicate that the D-GET method we developed surpasses existing methods on a custom dataset. The D-GET model, developed for DR classification using FFA images, significantly improves the detection of small-scale lesions. This advancement enhances the diagnosis and treatment of DR, establishing a solid foundation for its broader application across various domains of ophthalmic and general medical imaging.

Diabetic retinopathy (DR), a chronic disease in which the retina is damaged due to small vessel damage caused by diabetes mellitus, is one of the leading causes of vision impairment in diabetic patients. Detection of the earliest clinical sign of the advent of DR is a critical requirement for intervention and effective treatment. Ophthalmologists are trained to identify DR, based on examining specific minute changes in the eye - microaneurysms, retinal haemorrhages, macular edema and changes in the retinal blood vessels. Segmentation of microaneurysms (MA) is a critical requirement for the early diagnosis of DR and has been the primary focus of the research community over the past few years. In this work, a systematic review of existing literature is carried out to examine the diagnostic use of automated MA detection and segmentation for early DR diagnosis. We mainly focus on existing early DR diagnosis techniques to understand their strengths and weaknesses. Though early diagnosis is performed using colour fundus photography, fluorescein angiography or optical coherence tomography angiography images, our study is limited to colour fundus based techniques. The early DR diagnosis methodologies reviewed in this article can be broadly classified into classical image processing, conventional machine learning (ML), and deep learning (DL) based techniques. Though significant progress has been achieved in these three classes of early DR diagnosis, several challenges and gaps still exist, underscoring a considerable scope for the development of fully automated, user-friendly early DR diagnosis and grading systems. We discuss in detail the challenges that need to be addressed in designing such effective, efficient, and robust algorithms for early DR diagnosis systems and also the ample scope for future research in this area.

Optical coherence tomography angiography (OCTA) is a noninvasive imaging method that can be used for the staging of diabetic retinopathy. In addition, alterations in OCTA parameters can precede the clinical fundus changes. In this review, we aimed to assess the accuracy of OCTA in diagnosis and staging of diabetic retinopathy. Two independent reviewers participated in the literature search using electronic databases (PubMed, Embase, Cochrane Library Central Register of Controlled Trials, ISI, and Scopus) from inception till December 2020. The heterogeneity of data was assessed by Q statistics, Chi-square test and I2 index. Forty-four articles published from 2015 to the end of 2020 were included in this meta-analysis. Of these, 27 were case-control studies, 9 were case series, and 8 were cohort studies. In total, 4284 eyes of 3553 patients were assessed in this study. OCTA could differentiate diabetic retinopathy from diabetes without diabetic retinopathy with a sensitivity of 88% (95% CI: 85% to 92%) and specificity of 88% (95% CI: 85% to 91%). In addition, it could differentiate proliferative diabetic retinopathy from non-proliferative diabetic retinopathy with a sensitivity of 91% (95% CI: 86% to 95%) and specificity of 91% (95% CI:86% to 96%). The sensitivity of OCTA for diagnosing diabetic retinopathy was increased by the size of scan (3 × 3 mm: 85%; 6 × 6 mm: 91%, 12 × 12 mm: 96%). OCTA, as a non-invasive method, has acceptable sensitivity and specificity for diagnosis and classification of diabetic retinopathy. A larger scan size is associated with more sensitivity for discriminating diabetic retinopathy.

Diabetic retinopathy (DR) is an eye disorder triggered by diabetes that might result in loss of sight. Earlier diagnosis of DR is critical since it might cause loss of sight. Manual diagnoses of DR severity by ophthalmologists are time-consuming and challenging. As a result, there has been considerable attention on designing an automatic technique for DR detection using fundus photographs. In medical science, prognosis and diagnosis are the most challenging tasks due to the presence of fuzziness in medical images and the restricted subjectivity of the experts. Neutrosophic Set (NS) in medical image analysis provides an understanding of the NS concepts, together with knowledge of how to collect, handle, interpret, and analyze clinical images using NS techniques. The neutrosophic set (NS), which is a generality of fuzzy set, provides the overcoming prospect of the restriction of fuzzy-based models for the analysis of medical images. This manuscript develops a Robust Diabetic Retinopathy Detection and Grading using Neutrosophic Topological Vector Space (DRDG-NSTVS) technique on fundus images. The DRDG-NSTVS technique begins with Median Filter (MF) noise removal to optimize the clarity of fundus photographs by successfully eliminating noises. Later, the InceptionV3 is used to perform feature extraction for identifying complicated features and patterns related to DR. The parameter tuning is performed by the moth flame optimization (MFO) technique to ensure superior performance of the model. The final diagnoses and classification of DR are accomplished utilizing the NSTVS classifiers that easily perform the uncertainties inherent in medicinal statistics. The simulation was conducted on a benchmark dataset to examine the proposed model performance. This combined method gives a greatly reliable and accurate solution for the earlier diagnosis and detection of DR

Diabetic Retinopathy (DR) is a significant vision-threatening condition in diabetic individuals, mostly caused by damage to the retinal blood vessels. Microaneurysms, small red patches on the retina, are among the first indicators of DR. If left untreated, they can lead to serious vision problems or even blindness. Consequently, early and precise identification of microaneurysms is essential for efficient treatment and the prevention of disease progression. This research introduces an innovative deep learning framework for the automatic identification of microaneurysms in retinal fundus images with a Frozen-based VGG19 model. The suggested method has two main steps: pre-processing and classification. During preprocessing, lighting changes are corrected, and image scaling is applied to facilitate feature extraction. The modified Frozen-based VGG19 model is used in the classification phase to quickly identify microaneurysms without segmenting vessels. We tested the proposed technique on three benchmark datasets: E-ophtha, ROC, and DIARETDB1. We also compared it with existing transfer learning architectures such as VGG16 and VGG19. The experimental results show that the proposed model achieves 93.4% detection accuracy on the E-ophtha dataset, 88.3% on the ROC dataset, and 83.5% on the DIARETDB1 dataset. The work emphasises the capabilities of sophisticated deep learning models for early diagnosis of diabetic retinopathy, facilitating prompt clinical intervention and improving patient outcomes.

Diabetic retinopathy (DR) is an essential factor that has caused vision loss and even blindness in middle-aged and older adults. A system that can automatically perform DR diagnosis can help ophthalmologists save a lot of tedious work, such as DR grading or lesion detection. At the same time, patients can find their diseases earlier and perform the correct treatment. However, most of the existing methods require many DR annotations to train the model, and the DR data will vary to different degrees due to various shooting tools. The above problems lead to the inefficient use of existing data in the experiment, limiting actual deployment. To alleviate this problem, we propose a novel Graph Adversarial Transfer Learning (GATL) for DR diagnosis in a deep model through transfer learning, including intra-domain alignment and inter-domain alignment. The proposed GATL enjoys several merits. First, our GATL adopts the self-supervised training to save the annotating cost in the target domain thus this domain adaptation method can significantly reduce annotation cost compared to the supervised approaches. Second, we introduce the graph neural network to extract potential features between unknown samples. Third, to enhance the robustness of the model, we use adversarial training to perform both inter-domain and intra-domain alignment to further improve the model’s classification accuracy. GATL achieved 94.3%, 97.5%, and 91.1% in accuracy, sensitivity, and specificity in the APTOS dataset and 92.7%, 95.7%, and 89.7% in the EyePACS dataset, respectively. Extensive experimental results on two challenging benchmarks, including APTOS 2019 and EyePACS, demonstrate that the proposed GATL performs favorably against baseline DR classification methods.

Diabetic retinopathy is so imperceptible in its early stages that it can lead to blindness if it is not treated at the right time. If early and timely intervention, it will reduce visual impairment. Color doppler ultrasound imaging, fundus photography, Optical Tomography and Fundus Fluorescein angiography are commonly used in the diagnosis and treatment of diabetic retinopathy. Recently, deep learning techniques have been developed to assist in the early screening and diagnosis of diabetic retinas on a large scale. This article mainly introduces the research value and significance of deep learning in diabetic retina. In this paper, some deep learning and traditional machine learning methods for diabetic retinal image detection and severity classification are summarized, and some unresolved problems and prospects are discussed.

BackgroundAccurate segmentation of the foveal avascular zone (FAZ) is valuable for retinal imaging, as FAZ alterations are key biomarkers for diabetic retinopathy (DR). This study presents an automated framework exploring the feasibility of FAZ segmentation and DR classification using optical coherence tomography angiography (OCTA) images.MethodsIn this cross-sectional study conducted at Farabi Eye Hospital, Tehran, Iran, a two-step deep learning pipeline was developed. First, a neural network combining DeepLabv3+, EfficientNetB0, Squeeze-and-Excitation (SE) blocks, and Atrous Spatial Pyramid Pooling (ASPP) was trained to segment the FAZ from superficial capillary plexus (SCP) and deep capillary plexus (DCP) OCTA slabs. Second, a GoogLeNet-based convolutional neural network (CNN) classified segmented FAZ images into binary (normal vs. DR) and three-class (normal, non-proliferative DR [NPDR], proliferative DR [PDR]) categories to differentiate DR stages based on FAZ shape characteristics. For the classification task using the deep learning-generated segmented FAZ images as input, the data was split into 70% training, 10% validation, and 20% testing, with 5-fold cross-validation to mitigate overfitting. Data augmentation and Synthetic Minority Oversampling Technique (SMOTE) were applied to improve classification performance.ResultsThe final dataset comprised 253 OCTA scans (126 SCP, 127 DCP) from 161 eyes of 161 participants (one eye per participant), with 39 normal participants (24.2%), 78 patients with NPDR (48.4%), and 44 with PDR (27.3%). The mean age was 49.7 ± 11.8 years, and 54% were male. The FAZ segmentation network achieved a Dice similarity coefficient (DSC) of 97.5% across the dataset, achieving high precision even in challenging images. The classification model, using the deep learning generated segmented FAZ images as input, reached an area under the curve (AUC) of 100% for binary classification (normal vs. DR) and 87% for three-class classification (normal, NPDR, PDR) with oversampling.ConclusionThis system, with its potential for integrating into clinical workflows, offers a promising assistive tool for clinicians, which could enable earlier and more accurate diagnosis of diabetic retinopathy from OCTA images.Clinical trial numberNot applicable.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12886-025-04473-2.

Diabetic retinopathy (DR) is one of the most threatening complications in diabetic patients, leading to permanent blindness without timely treatment. However, DR screening is not only a time-consuming task that requires experienced ophthalmologists but also easy to produce misdiagnosis. In recent years, deep learning techniques based on convolutional neural networks have attracted increasing research attention in medical image analysis, especially for DR diagnosis. However, dataset labeling is expensive work and it is necessary for existing deep-learning-based DR detection models. For this study, a novel domain adaptation method (multi-model domain adaptation) is developed for unsupervised DR classification in unlabeled retinal images. At the same time, it only exploits discriminative information from multiple source models without access to any data. In detail, we integrate a weight mechanism into the multi-model-based domain adaptation by measuring the importance of each source domain in a novel way, and a weighted pseudo-labeling strategy is attached to the source feature extractors for training the target DR classification model. Extensive experiments are performed on four source datasets (DDR, IDRiD, Messidor, and Messidor-2) to a target domain APTOS 2019, showing that MMDA produces competitive performance for present state-of-the-art methods for DR classification. As a novel DR detection approach, this article presents a new domain adaptation solution for medical image analysis when the source data is unavailable.

Diabetic retinopathy (DR) is a late microvascular complication of Diabetes Mellitus (DM) that could lead to permanent blindness in patients, without early detection. Although adequate management of DM via regular eye examination can preserve vision in in 98% of the DR cases, DR screening and diagnoses based on clinical lesion features devised by expert clinicians; are costly, time-consuming and not sufficiently accurate. This raises the requirements for Artificial Intelligent (AI) systems which can accurately detect DR automatically and thus preventing DR before affecting vision. Hence, such systems can help clinician experts in certain cases and aid ophthalmologists in rapid diagnoses. To address such requirements, several approaches have been proposed in the literature that use Machine Learning (ML) and Deep Learning (DL) techniques to develop such systems. However, these approaches ignore the highly valuable clinical lesion features that could contribute significantly to the accurate detection of DR. Therefore, in this study we introduce a framework called DR-detector that employs the Extreme Gradient Boosting (XGBoost) ML model trained via the combination of the features extracted by the pretrained convolutional neural networks commonly known as transfer learning (TL) models and the clinical retinal lesion features for accurate detection of DR. The retinal lesion features are extracted via image segmentation technique using the UNET DL model and captures exudates (EXs), microaneurysms (MAs), and hemorrhages (HEMs) that are relevant lesions for DR detection. The feature combination approach implemented in DR-detector has been applied to two common TL models in the literature namely VGG-16 and ResNet-50. We trained the DR-detector model using a training dataset comprising of 1,840 color fundus images collected from e-ophtha, retinal lesions and APTOS 2019 Kaggle datasets of which 920 images are healthy. To validate the DR-detector model, we test the model on external dataset that consists of 81 healthy images collected from High-Resolution Fundus (HRF) dataset and MESSIDOR-2 datasets and 81 images with DR signs collected from Indian Diabetic Retinopathy Image Dataset (IDRID) dataset annotated for DR by expert. The experimental results show that the DR-detector model achieves a testing accuracy of 100% in detecting DR after training it with the combination of ResNet-50 and lesion features and 99.38% accuracy after training it with the combination of VGG-16 and lesion features. More importantly, the results also show a higher contribution of specific lesion features toward the performance of the DR-detector model. For instance, using only the hemorrhages feature to train the model, our model achieves an accuracy of 99.38 in detecting DR, which is higher than the accuracy when training the model with the combination of all lesion features (89%) and equal to the accuracy when training the model with the combination of all lesions and VGG-16 features together. This highlights the possibility of using only the clinical features, such as lesions that are clinically interpretable, to build the next generation of robust artificial intelligence (AI) systems with great clinical interpretability for DR detection. The code of the DR-detector framework is available on GitHub at https://github.com/Janga-Lab/DR-detector and can be readily employed for detecting DR from retinal image datasets.

Chapter 3 - Classification of Diabetic Retinopathy

Diabetic retinopathy (DR) is one of the leading causes of vision loss in adults and is one of the detrimental side effects of the mass prevalence of Diabetes Mellitus (DM). It is crucial to have an efficient screening method for early diagnosis of DR to prevent vision loss. This paper compares and analyzes the various Machine Learning (ML) techniques, from traditional ML to advanced Deep Learning models. We compared and analyzed the efficacy of Convolutional Neural Networks (CNNs), Capsule Networks (CapsNet), K-Nearest Neighbor (KNN), Support Vector Machine (SVM), decision trees, and Random Forests. This paper also considers determining factors in the evaluation, including contrast enhancements, noise reduction, grayscaling, etc We analyze recent research studies and compare methodologies and metrics, including accuracy, precision, sensitivity, and specificity. The findings highlight the advanced performance of Deep Learning (DL) models, with CapsNet achieving a remarkable accuracy of up to 97.98% and a high precision rate, outperforming other traditional ML methods. The Contrast Limited Adaptive Histogram Equalization (CLAHE) preprocessing technique substantially enhanced the model’s efficiency. Each ML method’s computational requirements are also considered. While most advanced deep learning methods performed better according to the metrics, they are more computationally complex, requiring more resources and data input. We also discussed how datasets like MESSIDOR could be more straightforward and contribute to highly evaluated performance and that there is a lack of consistency regarding benchmark datasets across papers in the field. Using the DL models facilitates accurate early detection for DR screening, can potentially reduce vision loss risks, and improves accessibility and cost-efficiency of eye screening. Further research is recommended to extend our findings by building models with public datasets, experimenting with ensembles of DL and traditional ML models, and considering testing high-performing models like CapsNet.

This study introduces a framework aimed at bolstering the effectiveness of deep learning models in image classification tasks, with a specialized focus on diabetic retinopathy detection. The proposed strategy employs transfer and multitask learning methods for the consistent extraction of insights and refinement of pre-established models. The capability of multitask learning to simultaneously acquire knowledge from numerous related tasks, leveraging shared information and inter-dependencies, leads to considerable performance boosts. Integrated into this process is the feature guideline learning task, which empowers the model to have an attention to abnormal areas. This task is initially trained on a small collection of labeled images to develop a custom pre-trained model, which later forms the basis for subsequent classification tasks. The effectiveness of this proposed approach is validated through tests on various data sets. It has been proven that our method contributes considerable advancements in image classification performance. This research offers valuable perspectives on the detection and classification of diabetic retinopathy, providing solutions to problems originating from limited data, and shows promising potential for improving patient care.