A Comparative Study of Deep Learning and Transfer Learning in Detection of Diabetic Retinopathy (original) (raw)
Related papers
A DEEP TRANSFER LEARNING FRAMEWORK for the STAGING of DIABETIC RETINOPATHY
DergiPark (Istanbul University), 2022
Diabetes is a highly prevalent and increasingly common health disorder, resulting in health complications such as vision loss. Diabetic retinopathy (DR) is the most common form of diabetescaused eye disease. Early diagnosis and treatment are crucial to prevent vision loss. DR is a progressive disease composed of five stages. The accurate diagnosis of DR stages is highly important in guiding the treatment process. In this study, we propose a deep transfer learning framework for automatic detection of DR stages. We examine our proposed model by comparing different convolutional neural networks architectures: VGGNet19, DenseNet201, and ResNet152. Our results demonstrate better accuracy after applying transfer learning and hyper-parameter tuning to classify the fundus images. When the general test accuracy and the performance evaluations are compared, the DenseNet201 model is observed with the highest test accuracy of 82.7%. Among the classification algorithms, the highest AUC value is 94.1% obtained with RestNet152.
A transfer learning with deep neural network approach for diabetic retinopathy classification
International Journal of Electrical and Computer Engineering (IJECE), 2021
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 network (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.
1. We developed a novel predictive model based on Efficient Net b3 pre-trained Convolutional Neural Network for reliable classification of Diabetic Retinopathy into 5 Stages, namely: 0 – No DR, 1 – Mild DR, 2 – Moderate DR, 3 – Severe DR, 4 – Proliferate DR. These 5 stages are based on the severity of diabetic retinopathy. Since treatment varies based on the severity of diabetic retinopathy. So, if 5 stages are predicted appropriately then it helps doctor to diagnose the patient based on the stage. 2. We used three different pre-trained CNN models: VGG Net, Res Net (which are scaled by their depth to get higher versions), and Efficient Net (which uses an efficient compound scaling method that scales width, height, and resolution to get higher versions). We observed Efficient Net could classify the images into more than one class, unlike Res Net and VGG Net. The reason could be models like Res Net and VGG Net is not capable of extracting complex features from all the stages of diabet...
Transfer Learning based Detection of Diabetic Retinopathy from Small Dataset
2019
Annotated training data insufficiency remains to be one of the challenges of applying deep learning in medical data classification problems. Transfer learning from an already trained deep convolutional network can be used to reduce the cost of training from scratch and to train with small training data for deep learning. This raises the question of whether we can use transfer learning to overcome the training data insufficiency problem in deep learning based medical data classifications. Deep convolutional networks have been achieving high performance results on the ImageNet Large Scale Visual Recognition Competition (ILSVRC) image classification challenge. One example is the Inception-V3 model that was the first runner up on the ILSVRC 2015 challenge. Inception modules that help to extract different sized features of input images in one level of convolution are the unique features of the Inception-V3. In this work, we have used a pre-trained Inception-V3 model to take advantage of its Inception modules for Diabetic Retinopathy detection. In order to tackle the labelled data insufficiency problem, we subsampled a smaller version of the Kaggle Diabetic Retinopathy classification challenge dataset for model training, and tested the model's accuracy on a previously unseen data subset. Our technique could be used in other deep learning based medical image classification problems facing the challenge of labeled training data insufficiency.
AECE, 2021
Diabetic Retinopathy (DR) stays a main source of vision deterioration around world and it is getting exacerbated day by day. Almost no warning signs for detecting DR which will be greater challenge with us today. So, it is extremely preferred that DR has to be discovered on time. Adversely, the existing result involves an ophthalmologist to manually check and identify DR by positioning the exudates related with vascular irregularity due to diabetes from fundus image. In this work, we are able to classify images based on different severity levels through an automatic DR classification system. To extract specific features of image without any loss in spatial information, a Convolutional Neural Network (CNN) models which possesses an image with a distinct weight matrix is used. In the beginning, we estimate various CNN models to conclude the best performing CNN for DR classification with an objective to obtain much better accuracy. In the classification of DR disease with transfer learning using deep CNN models, 97.72% of accuracy is provided by the proposed CNN model for Kaggle dataset. The proposed CNN model provides a classification accuracy of 97.58% for MESSIDOR dataset. The proposed technique provides better results than other state-ofart methods.
Diabetic Retinopathy Classification using Transfer learning
International Journal of Advanced Trends in Computer Science and Engineering , 2023
Diabetic Retinopathy (DR) is an eye illness that impacts individuals who have diabetes and damages their retina over time, eventually causing blindness. Due to lesions in the retina that are formed because of retinal blood vessel rupture, it impairs vision and, in the worst-case scenario, results in severe blindness. To prevent severity and to lessen challenges in identifying tiny lesions throughout the disease's advanced stages, it is now crucial to diagnose the condition early as, it manifests itself without any symptoms. Even ophthalmologists find it challenging and time-consuming to identify this condition. Early DR case identification and classification is essential for delivering the required medical care. This study proposes applying deep learning techniques to detect DR in retinal fundus images. The data acquired for this process may be incomplete and imbalanced. Data augmentation balances the data and increase the quantity of retinal images. As deep-learning algorithms need more data to process, DCGAN Augmentation technique is employed. The CNN (Convolutional Neural Network) methods, specifically the VGG16 and DenseNet121 architectures, are employed for DR early detection in order to let patients to receive therapy at the appropriate time..
Reducing Overfitting in Diabetic Retinopathy Detection using Transfer Learning
2020 IEEE 5th International Conference on Computing Communication and Automation (ICCCA), 2020
Diabetic retinopathy is a diabetes complication which causes damage to the retina. It is highly prevalent in patients who have had diabetes for 20 years or more. Current deep learning methods of detection of this disease are not efficient with the respect to the number of images required for training and the transfer learning methods use pre-trained networks, which contains many features that are irrelevant to the original task data. In this paper, we pre-train our network in a self supervised manner on the training data, such that the network learns features relevant to the task data without overfitting. We propose a data efficient architecture based on autoencoder and VGG network and a method to train the network such that it is regularized by an autoencoder network to prevent overfitting and improve test data performance.
A deep transfer learning approach for identification of diabetic retinopathy using data augmentation
IAES International Journal of Artificial Intelligence, 2022
In ophthalmology, deep learning acts as a computer-based tool with numerous potential capabilities and efficacy. Throughout the world, diabetic retinopathy (DR) is considered as a principal cause of disease however loss of sight cannot be seen in adults aged 20-74 years. The primary objective for early detection of DR is screening on a regular basis at separate intervals which should have a time difference of every ten to twenty months for the patients with no or mild case of DR. Regular screening plays a major role to prevent vision loss, the expected cases increase from 415 million in 2015 to 642 million in 2040 means is a challenging task of ophthalmologists to do screening and follow-up representations. In this research, a transfer learning model was proposed with data augmentation techniques and gaussian-blur, circle-crop pre-processing techniques combination to identify every stage of DR using Resnet 50 with top layers. Models are prepared with Kaggle Asia Pacific Tele-Ophthalmology Society blindness dataset on a top line graphical processing data. The result depicts- the comparison of classification metrics using synthetic and non-synthetic images and achieve accuracy of 91% using the synthetic data and 86% accuracy without using synthetic data.
Transfer Learning-Based Model for Diabetic Retinopathy Diagnosis Using Retinal Images
Brain Sciences
Diabetic retinopathy (DR) is a visual obstacle caused by diabetic disease, which forms because of long-standing diabetes mellitus, which damages the retinal blood vessels. This disease is considered one of the principal causes of sightlessness and accounts for more than 158 million cases all over the world. Since early detection and classification could diminish the visual impairment, it is significant to develop an automated DR diagnosis method. Although deep learning models provide automatic feature extraction and classification, training such models from scratch requires a larger annotated dataset. The availability of annotated training datasets is considered a core issue for implementing deep learning in the classification of medical images. The models based on transfer learning are widely adopted by the researchers to overcome annotated data insufficiency problems and computational overhead. In the proposed study, features are extracted from fundus images using the pre-trained ...
Diabetic retinopathy stage detection using convolutional fine-tuned transfer Learning model
International Journal of Experimental Research and Review, 2023
Diabetic Retinopathy (DR) is a prevalent eye condition that occurs as a frequent complication among individuals with diabetes, particularly those who have been living with the disease for an extended period of time. This study uses fundus images to diagnose DR at five stages from early to late with No DR, Mild, Moderate, Severe, and Proliferative DR, commonly known as Stage 0 to Stage 4, respectively. This will aid in the timely treatment of diabetic patients preventing them from developing DR as early as possible. We used two most popular open-source datasets, the DR Detection database, namely APTOS 2019 and EyePACS, and combined them to create a larger dataset to trade off the data-centric obstacle and shortfall for any Deep Learning-based prediction models. Data augmentation and preprocessing techniques are applied to the images before feeding them to the proposed model to get a more accurate and efficient one. In the modern age oriented to Artificial Intelligence (AI), it is necessary to thoroughly analyze the identification of DR based on the existing Deep Learning (DL) models. After learning about the limitations of existing models, we have fine-tuned the ResNet50, DenseNet201 and InceptionV3 to enhance the model performance of the detection and categorization of DR. We have since proposed three Deep Convolutional Neural Networks (DCNN) models with better outcome based on accuracy than the existing state-of-the-art (SOTA) models. The fine-tuned DenseNet201 model, among the other two, performed significantly better with a validation accuracy of 90.04% and a negligible amount of loss, irrespective of each class, under the best configurable test conditions.