Effective Deep Learning Data Augmentation Techniques for Diabetic Retinopathy Classification

Quadratic Weighted Kappa (QWK) is a statistic to measure the agreement between two annotators. QWK has been widely used as the evaluation measure for various medical imaging problems, where, the class labels have a natural ordering, e.g., no Diabetic Retinopathy (DR), mild DR, and severe DR. The easiest way to treat the classification problem with these ordinal labels is to consider the problem as a multiclass classification problem and apply the Cross Entropy (CE) loss. However, when applying CE loss the order of the classes becomes meaningless, i.e., the loss will be same if a healthy image is classified into mild DR or severe DR. At the same time, the QWK score will be severely affected if a healthy image is classified into severe DR than mild DR. The most appropriate way to get a better classification score is to optimize the evaluation measure itself. i.e., directly optimize the QWK statistics. However, this optimization may hinder the learning, and may lead to sub-optimal solutions, and therefore, may give lower performance than expected. On the other hand, Ordinal Regression (OR) based approaches also can be used for such problems. The main focus of this work is to investigate which loss function (CE loss, QWK loss or OR loss) is the most appropriate one to the Convolutional Neural Network-based ordinal classification problems, where, QWK is used as the evaluation measure. Experiments on two public datasets, Diabetic Retinopathy and Prostate Cancer, with two different network architectures suggest that directly optimizing QWK is the better choice when small networks are used. On the other hand, we found that for large networks OR based loss function gives better performance.

Diabetic Retinopathy (DR), chronic progressive disease of the eye, may give rise to permanent sight loss. Clinicians use fundus pictures to check if DR is present and rely on physicians to diagnose the stage or severity by visual inspection of the images. In relying on a clinician’s subjective prognosis, this is deemed a procedure that takes a lot of time and susceptible to misjudgements. In discovering DR, poor Quadratic Weighted Kappa (QWK) scores have resulted from poor quality of pictures and imbalanced distribution of classes. Even though studies have shown high accuracy, sensitivity, specificity and ROC metrics, their limitation is that they do not consider the level of disparity across the classified labels. The QWK score demonstrates that even if an algorithm presents high accuracy, it is still not best fit to classify DR into its 5 classes. Many researchers have tried fine-tuning the neural network to create noise-resistant deep learning and recorded high accuracy and sensitivity but low QWK scores. The problem with the other methods is mainly pre-processing of the images and model building patterns. Most of the studied literature lacks the image augmentation step which might lead to an erroneous result. This research aims to create an algorithm from deep learning models with a data augmentation step and demonstrate how important it is for attaining better QWK scores for all stages of diabetic retinopathy. The model met study objectives and obtained an accuracy of 93% and a QWK score of 0.961. The outcomes show that the method can make accurate predictions without need for human feature extraction and that it may be used as an early DR diagnostic and staging screening tool.

Psoriasis is a serious autoimmune skin disorder. It appears up on the skin as red, scaly regions. This effect on people lives who impacted with this disease. Early and accurate detection is more important for recovering from this disease. This method provides a new way to detect psoriasis disease. It utilises image augmentation and folder division techniques for improving the performance of deep learning models. The image augmentation techniques like as rotation, scaling, flipping, and color shifting were utilised on a set of psoriasis images. This avoided the models from refitting and improved the variety of the dataset. The folder division technique helped to organize training, validation, and testing datasets. This structure improves the deep learning model's applicability to different psoriasis symptoms. The results show that combining image augmentation with folder division techniques improves the accuracy and early detection of psoriasis. Dermatologists can use this as a useful tool for diagnosing skin problems

Diabetic Retinopathy (DR) is a significant eye disease, which is caused by the damage of retina. To provide the best timely treatment, it is necessary to detect DR in early stages. Firstly, advanced sequential image preprocessing and segmentation techniques are employed for accurate localizing and isolating the affected regions in retinal images. Secondly, a voting ensemble classifier is introduced using a deep neural network model, which combines the predictions of multiple CNN models i.e., ResNet50, VGG16, VGG19 and GoogLeNet to enhance the overall classification performance of the proposed model. Our proposed model, named VEnDR-Net (Voting Ensemble for Diabetic Retinopathy classification using deep neural networks), implements on the EyePACS dataset and achieves 0.97 sensitivity, 0.97 specificity, 0.98 accuracy, 0.98 precision, and 0.97 F1- Score, respectively. The enhancement of the performance is 1.49% in accuracy over the other existing models. Lastly, the research addresses the grading of diabetic retinopathy by aligning the classification results with a standardized grading system, providing clinicians with accurate severity assessment for effective treatment decision.

Diabetic retinopathy stands as a leading cause of blindness among people. Manual examination of DR images is labor-intensive and prone to error. Existing methods to detect this disease often rely on handcrafted features which limit the adaptability and classification accuracy. Thus, the aim of this research is to develop an automated and efficient system for early detection and accurate grading of diabetic retinopathy severity with less time consumption. In our research, we have developed a deep neural network named RSG-Net (Retinopathy Severity Grading) to classify DR into 4 stages (multi-class classification) and 2 stages (binary classification). The dataset utilized in this study is Messidor-1. In preprocessing, we have used Histogram Equalization to improve image contrast and denoising techniques to remove noise and artifacts which enhanced the clarity of the fundus images. We applied data augmentation techniques to our preprocessed images in order to tackle class imbalance issues. Augmentation techniques involve flipping, rotation, zooming and adjustment of color, contrast and brightness. The proposed RSG-Net model contains convolutional layers to perform automatic feature extraction from the input images and batch normalization layers to improve training speed and performance. The model also contains max pooling, drop out and fully connected layers. Our proposed RSG-Net model achieved a testing accuracy of 99.36%, specificity of 99.79% and a sensitivity of 99.41% in classifying diabetic retinopathy into 4 grades and it achieved 99.37% accuracy, 100% sensitivity and 98.62% specificity in classifying DR into 2 grades. The performance of RSG-Net is also compared with other state-of-the-art methodologies where it outperformed these methods.

To develop a deep learning (DL) model to detect morphologic patterns of diabetic macular edema (DME) based on optical coherence tomography (OCT) images. In the training set, 12,365 OCT images were extracted from a public data set and an ophthalmic center. A total of 656 OCT images were extracted from another ophthalmic center for external validation. The presence or absence of three OCT patterns of DME, including diffused retinal thickening, cystoid macular edema, and serous retinal detachment, was labeled with 1 or 0, respectively. A DL model was trained to detect three OCT patterns of DME. The occlusion test was applied for the visualization of the DL model. Applying 5-fold cross-validation method in internal validation, the area under the receiver operating characteristic curve for the detection of three OCT patterns (i.e., diffused retinal thickening, cystoid macular edema, and serous retinal detachment) was 0.971, 0.974, and 0.994, respectively, with an accuracy of 93.0%, 95.1%, and 98.8%, respectively, a sensitivity of 93.5%, 94.5%, and 96.7%, respectively, and a specificity of 92.3%, 95.6%, and 99.3%, respectively. In external validation, the area under the receiver operating characteristic curve was 0.970, 0.997, and 0.997, respectively, with an accuracy of 90.2%, 95.4%, and 95.9%, respectively, a sensitivity of 80.1%, 93.4%, and 94.9%, respectively, and a specificity of 97.6%, 97.2%, and 96.5%, respectively. The occlusion test showed that the DL model could successfully identify the pathologic regions most critical for detection. Our DL model demonstrated high accuracy and transparency in the detection of OCT patterns of DME. These results emphasized the potential of artificial intelligence in assisting clinical decision-making processes in patients with DME.

Prediction of the stage of cancer plays an important role in planning the course of treatment and has been largely reliant on imaging tools which do not capture molecular events that cause cancer progression. Gene-expression data-based analyses are able to identify these events, allowing RNA-sequence and microarray cancer data to be used for cancer analyses. Breast cancer is the most common cancer worldwide, and is classified into four stages - stages 1, 2, 3, and 4 [2]. While machine learning models have previously been explored to perform stage classification with limited success, multi-class stage classification has not had significant progress. There is a need for improved multi-class classification models, such as by investigating deep learning models. Gene-expression-based cancer data is characterised by the small size of available datasets, class imbalance, and high dimensionality. Class balancing methods must be applied to the dataset. Since all the genes are not necessary for stage prediction, retaining only the necessary genes can improve classification accuracy. The breast cancer samples are to be classified into 4 classes of stages 1 to 4. Invasive ductal carcinoma breast cancer samples are obtained from The Cancer Genome Atlas (TCGA) and Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) datasets and combined. Two class balancing techniques are explored, synthetic minority oversampling technique (SMOTE) and SMOTE followed by random undersampling. A hybrid feature selection pipeline is proposed, with three pipelines explored involving combinations of filter and embedded feature selection methods: Pipeline 1 - minimum-redundancy maximum-relevancy (mRMR) and correlation feature selection (CFS), Pipeline 2 - mRMR, mutual information (MI) and CFS, and Pipeline 3 - mRMR and support vector machine-recursive feature elimination (SVM-RFE). The classification is done using deep learning models, namely deep neural network, convolutional neural network, recurrent neural network, a modified deep neural network, and an AutoKeras generated model. Classification performance post class-balancing and various feature selection techniques show marked improvement over classification prior to feature selection. The best multiclass classification was found to be by a deep neural network post SMOTE and random undersampling, and feature selection using mRMR and recursive feature elimination, with a Cohen-Kappa score of 0.303 and a classification accuracy of 53.1%. For binary classification into early and late-stage cancer, the best performance is obtained by a modified deep neural network (DNN) post SMOTE and random undersampling, and feature selection using mRMR and recursive feature elimination, with an accuracy of 81.0% and a Cohen-Kappa score (CKS) of 0.280. This pipeline also showed improved multiclass classification performance on neuroblastoma cancer data, with a best area under the receiver operating characteristic (auROC) curve score of 0.872, as compared to 0.71 obtained in previous work, an improvement of 22.81%. The results and analysis reveal that feature selection techniques play a vital role in gene-expression data-based classification, and the proposed hybrid feature selection pipeline improves classification performance. Multi-class classification is possible using deep learning models, though further improvement particularly in late-stage classification is necessary and should be explored further.

Background Diabetic Retinopathy (DR) remains a leading cause of preventable blindness, with rising prevalence particularly in low-resource settings. Traditional DR screening relies on specialist ophthalmologic assessment, posing barriers to early detection. Recent advances in deep learning have enabled the development of automated tools that may enhance the scalability and accessibility of DR screening. Objectives To compare the performance of two convolutional neural network (CNN) architectures-EfficientNet-B5 and ResNet-50-for automated DR detection using publicly available retinal fundus image datasets. Methods We utilized a transfer learning approach on two large datasets: EyePACS 2015 (n = 35,126) for training/validation and APTOS 2019 (n = 3,662) for external testing. Images were preprocessed with resizing, cropping, and contrast enhancement. The EfficientNet-B5 and ResNet-50 models were independently trained and evaluated using key metrics: accuracy, sensitivity, specificity, area under the ROC curve (AUC), and Quadratic Weighted Kappa (QWK), with classification across DR severity levels. Results EfficientNet-B5 achieved a QWK of 92%, sensitivity of 98%, specificity of 93%, and AUC of 96% on the test dataset. ResNet-50 demonstrated similarly high performance with an accuracy of 96%, QWK of 86%, sensitivity of 94%, specificity of 97%, and AUC of 96%. Both models showed robust generalization across severity grades, indicating strong discriminative capabilities. Conclusions The results highlighted that our models are viable architectures for AI-driven DR screening, with EfficientNet-B5 showing slightly higher agreement scores and ResNet-50 offering stronger specificity. These models hold promise for deployment in DR screening programs, especially in resource-limited settings where ophthalmologic expertise is scarce. Future work should focus on clinical validation and integration into public health systems to improve early DR detection and reduce vision-related morbidity. Key messages • Integrating artificial intelligence (AI) into diabetic retinopathy screening can enhance early detection and support equitable access to eye care, particularly in resource-limited settings. • AI-based screening tools offer a scalable and effective approach to strengthen public health strategies aimed at preventing vision loss from diabetic complications.

Diabetic Retinopathy (DR) is one of the severe microvascular conditions that affect the retina of the eye. It can result in permanent partial or total blindness if not discovered and treated early. Across the world, diabetes impacts in excess of 50% of people below 70 years of age. Many diabetic people, however, fail to notice the disease and suffer visual impairments as a result of the time it takes to see an ophthalmologist who screens and analyses the patient's retina. This research study focuses on the automated detection and categorization of DR based on distinct severity levels using fundus images. The method starts with a pre-processing step to get rid of any extraneous noise from the edges and aids in the focus on the area of interest. The DR fundus images were then categorized into various severity levels using DL models like VGG19, ResNet50, and Inception-V3 Deep Learning (DL) models. The logic for the presented DL model has been tested on the EyePACS (kaggle) DR dataset. In terms of classification accuracy, the suggested DL model with ResNet50 configuration surpassed other models such as VGG19 and Inception-V3 as well as other current models, according to the findings of the research.

Diabetic retinopathy is an eye disease caused by high blood sugar and pressure which damages the blood vessels in the eye. Diabetic retinopathy is the root cause of more than 1% of the blindness worldwide. Early detection of this disease is crucial as it prevents it from progressing to a more severe level. However, the current machine learning-based approaches for detecting the severity level of diabetic retinopathy are either, i) rely on manually extracting features which makes an approach unpractical, or ii) trained on small dataset thus cannot be generalized. In this study, we propose a transfer learning-based approach for detecting the severity level of the diabetic retinopathy with high accuracy. Our model is a deep learning model based on global average pooling (GAP) technique with various pre-trained convolutional neural net- work (CNN) models. The experimental results of our approach, in which our best model achieved 82.4% quadratic weighted kappa (QWK), corroborate the ability of our model to detect the severity level of diabetic retinopathy efficiently.

Diabetic retinopathy (DR) is a severe complication of diabetes. It affects a large portion of the population of the Kingdom of Saudi Arabia. Existing systems assist clinicians in treating DR patients. However, these systems entail significantly high computational costs. In addition, dataset imbalances may lead existing DR detection systems to produce false positive outcomes. Therefore, the author intended to develop a lightweight deep-learning (DL)-based DR-severity grading system that could be used with limited computational resources. The proposed model followed an image pre-processing approach to overcome the noise and artifacts found in fundus images. A feature extraction process using the You Only Look Once (Yolo) V7 technique was suggested. It was used to provide feature sets. The author employed a tailored quantum marine predator algorithm (QMPA) for selecting appropriate features. A hyperparameter-optimized MobileNet V3 model was utilized for predicting severity levels using images. The author generalized the proposed model using the APTOS and EyePacs datasets. The APTOS dataset contained 5590 fundus images, whereas the EyePacs dataset included 35,100 images. The outcome of the comparative analysis revealed that the proposed model achieved an accuracy of 98.0 and 98.4 and an F1 Score of 93.7 and 93.1 in the APTOS and EyePacs datasets, respectively. In terms of computational complexity, the proposed DR model required fewer parameters, fewer floating-point operations (FLOPs), a lower learning rate, and less training time to learn the key patterns of the fundus images. The lightweight nature of the proposed model can allow healthcare centers to serve patients in remote locations. The proposed model can be implemented as a mobile application to support clinicians in treating DR patients. In the future, the author will focus on improving the proposed model's efficiency to detect DR from low-quality fundus images.

Application of deep learning image assessment software VeriSee™ for diabetic retinopathy screening

Diabetic retinopathy (DR) is an eye issue of prolonged diabetes, which continues to be a major cause of vision loss among middle-aged adults. Early detection of DR through advanced imaging and artificial intelligence (AI) techniques can lead to a significant reduction in risk and severity of visual loss. This proposed ARMDiaRD combines state-of-the-art techniques from EfficientNet, Swin Transformers, and multi-scale feature fusion (MSFF) to enhance multiclass classification of DR severity levels across diverse generalised datasets. In the proposed architecture, a global context (GC) block is integrated into EfficientNet to capture long-range dependencies and contextual relationships. This is followed by Swin Transformer layers equipped with an MSFF block that hierarchically aggregates features from multiple levels, enabling the model to learn richer and more discriminative representations. The fundus images of four publicly available datasets for DR: asia pacific tele ophthalmology society (APTOS), Indian diabetic retinopathy image dataset (IDRID), MESSIDOR-V1, and EYEPACS, are combined through a comprehensive aggregation process to facilitate the training and testing of the generalised model. This design demonstrates consistent improvements across multiple evaluation metrics, underlining its potential to reduce misclassification in medical diagnosis. To evaluate the proposed model, the performance metrics, such as accuracy, precision, recall, specificity, and F1-score, are calculated along with quadratic weighted kappa (QWK), Spearman rank correlation coefficient, and mean absolute error (MAE). Simulation results revealed that the proposed model achieved an accuracy of 87.59%, precision of 87.6%, recall of 87.9%, QWK of 91.47% and Spearman coefficients with 92.53% respectively. Importantly, the MAE, a critical metric for evaluating false predictions in medical diagnosis, is 0.1736 in the proposed model. The outcome clearly demonstrates the supremacy of the proposed model over the other models in handling multiple larger datasets. This research shows the ubiquitous nature of the proposed model to predict the severity grades using larger complex datasets. This proposed ARMDiaRD confirms that combining local, global, and hierarchical characteristics is highly effective in preventing overfitting issues with the existing architecture, such as CNN, when tested using new data for reliable clinical image analysis.

BackgroundCytomegalovirus (CMV) reactivation in patients with severe ulcerative colitis (UC) leads to worse outcomes; yet, early detection remains challenging due to the reliance on time-intensive biopsy procedures.ObjectiveThis study explores the use of deep learning to differentiate CMV from severe UC through endoscopic imaging, offering a potential noninvasive diagnostic tool.MethodsWe analyzed 86 endoscopic images using an ensemble of deep learning models, including DenseNet (Densely Connected Convolutional Network) 121 pretrained on ImageNet. Advanced preprocessing and test-time augmentation (TTA) were applied to optimize model performance. The models were evaluated using metrics such as accuracy, precision, recall, F1-score, and area under the curve.ResultsThe ensemble approach, enhanced by TTA, achieved high performance, with an accuracy of 0.836, precision of 0.850, recall of 0.904, and an F1-score of 0.875. Models without TTA showed a significant drop in these metrics, emphasizing TTA’s importance in improving classification performance.ConclusionsThis study demonstrates that deep learning models can effectively distinguish CMV from severe UC in endoscopic images, paving the way for early, noninvasive diagnosis and improved patient care.

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.