A Deep Learning Ensemble Approach for Diabetic Retinopathy Detection

Qummar, Sehrish; Khan, Fiaz Gul; Shah, Sayyar Ali; Khan, Ahmad; Shamshirband, Shahaboddin; Rehman, Zia Ur; Jadoon, Waqas

doi:10.1109/access.2019.2947484

Cited by 355 publications

(153 citation statements)

References 25 publications

(33 reference statements)

Supporting

Mentioning

148

Contrasting

Unclassified

Order By: Relevance

“…The test information was isolated into preparing, cross approval, and testing purposes. In order to train the network, training and transfer function was chosen with a fix learning rate (Qummar et al, 2019). For preparing, the capacity LM‐quickest back proliferation calculation is profoundly suggested as a first‐decision directed calculation (Sadrzadeh, Mohammadi, Ivakpour, & Kasiri, 2008), despite the fact that it requires more memory than other calculations.…”

Section: Methodsmentioning

confidence: 99%

Comparative performance analysis of enzyme inactivation of soy milk by using RSM and ANN

Kumar

Rao

Rana

et al. 2020

J Food Process Engineering

View full text Add to dashboard Cite

The presence of antinutritional factors as trypsin inhibitors (TIA) and lipoxygenase (LOX) in soy milk produces indigestion and off‐flavor due to oxidation of linoleic acid to hyperoxide. The objective of the study was to determine the prediction capacity of response surface methodology (RSM) and artificial neural network (ANN) for enzyme inactivation of soy milk. The microwave and thermo‐sonication method were used to prepare and treat the sample. Statistical parameters like NRMSE and %MAE were used to compare and evaluate the final result. NRMSE value had shown five times better results in the case of ANN (0.015) compared to RSM (0.082). Similarly, the % MAE value was also fivefold better in the case of ANN. In the case of RSM, Chi‐square values for TIA and LOX were 317.32 and 146.73, respectively. Whereas for ANN, the value was 4.68 and 2.69, respectively. So, it can be concluded that the prediction capacity of ANN is better than RSM. Practical Applications This modeling has a tremendous potential contributing to prediction of enzymes inactivation without real‐time experiments with great accuracy. Moreover, a number of industries are producing soy milk nowadays, and it could be helpful in elimination of beany flavor with desirable inactivation of lipoxygenase and enhancing the digestibility by removing trypsin inhibitors with appropriate environmental conditions. In addition, it can also pave the way for a number of food processing technology in which enzymatic inactivation is prominent and the prediction of percentage inactivation is crucial. This technique could be applied to predict the yield of any process after maintaining a certain treatment and process condition. Process engineering of any industry and biochemical processes are very sensitive to small changes in parameters and their effects could be predicted by using the modern approach of artificial neural network with more preciseness without running the actual experiments.

show abstract

Section: Methodsmentioning

confidence: 99%

Comparative performance analysis of enzyme inactivation of soy milk by using RSM and ANN

Kumar

Rao

Rana

et al. 2020

J Food Process Engineering

View full text Add to dashboard Cite

show abstract

“…There have been several studies that have used the EyePACS dataset for training a DR AI [28,[44][45][46][47]. However, these studies are not directly comparable to the DRCNN presented here, as each study uses different (sometimes private) training sets, validation sets, DR grading schemes and performance reporting metrics.…”

Section: Sensitivity Upliftmentioning

confidence: 99%

Towards implementation of AI in New Zealand national diabetic screening program: Cloud-based, robust, and bespoke

et al. 2020

View full text Add to dashboard Cite

Convolutional Neural Networks (CNNs) have become a prominent method of AI implementation in medical classification tasks. Grading Diabetic Retinopathy (DR) has been at the forefront of the development of AI for ophthalmology. However, major obstacles remain in the generalization of these CNNs onto real-world DR screening programs. We believe these difficulties are due to use of 1) small training datasets (<5,000 images), 2) private and 'curated' repositories, 3) locally implemented CNN implementation methods, while 4) relying on measured Area Under the Curve (AUC) as the sole measure of CNN performance. To address these issues, the public EyePACS Kaggle Diabetic Retinopathy dataset was uploaded onto Microsoft Azure™ cloud platform. Two CNNs were trained; 1 a "Quality Assurance", and 2. a "Classifier". The Diabetic Retinopathy classifier CNN (DRCNN) performance was then tested both on 'un-curated' as well as the 'curated' test set created by the "Quality Assessment" CNN model. Finally, the sensitivity of the DRCNNs was boosted using two post-training techniques. Our DRCNN proved to be robust, as its performance was similar on 'curated' and 'un-curated' test sets. The implementation of 'cascading thresholds' and 'max margin' techniques led to significant improvements in the DRCNN's sensitivity, while also enhancing the specificity of other grades. OPEN ACCESSCitation: Xie L, Yang S, Squirrell D, Vaghefi E (2020) Towards implementation of AI in New Zealand national diabetic screening program: Cloud-based, robust, and bespoke. PLoS ONE 15 (4): e0225015. https://doi.org/10.

show abstract

“…In recent years, due to the explosive growth of image data on the Internet, deep learning has been widely used in image classification [5], [6]. For example, deep CNN based approaches to detecting DR are proposed in [7], [8] and [9]. To get high precision, such approaches usually use very deep networks and very large scale of datasets.…”

Section: Introductionmentioning

confidence: 99%

An Approach to Detecting Diabetic Retinopathy Based on Integrated Shallow Convolutional Neural Networks

Chen

Yang

et al. 2020

IEEE Access

View full text Add to dashboard Cite

The early detection of Diabetic Retinopathy (DR) is critical for diabetics to lower the blindness risks. Many studies represent that Deep Convolutional Neural Network (CNN) based approaches are effective to enable automatic DR detection through classifying retinal images of patients. Such approaches usually depend on a very large dataset composed of retinal images with predefined classification labels to support their CNN training. However, in some occasions, it is not so easy to get enough welllabelled images to act as model training samples. At the same time, when a CNN becomes deeper, its training will not only take much longer time, but also be more likely to lead to overfitting, especially on a large training dataset. Therefore, it is meaningful to explore a simpler CNN based approach that is still effective on small datasets to classify retinal images. In this paper, an approach to retinal image classification is proposed based on the integration of multi-scale shallow CNNs. Experiments on public datasets show that, on small datasets, the proposed approach can improve the classification accuracy by 3% compared with current representative integrated CNN learning approaches. On the bigger dataset, the proposed approach can improve the classification accuracy by 3% to 9% compared with other representative approaches such as traditional CNN, LCNN and VGG16noFC. The evaluation also represents that, though the classification accuracy of the proposed approach declines by 6% on the smallest dataset containing only 10% samples of the original dataset, its time cost declines to about 30% of that on the original dataset. INDEX TERMS convolutional neural network, diabetic retinopathy, image classification, integrated learning, performance integration.

show abstract

A Deep Learning Ensemble Approach for Diabetic Retinopathy Detection

Cited by 355 publications

References 25 publications

Comparative performance analysis of enzyme inactivation of soy milk by using RSM and ANN

Comparative performance analysis of enzyme inactivation of soy milk by using RSM and ANN

Towards implementation of AI in New Zealand national diabetic screening program: Cloud-based, robust, and bespoke

An Approach to Detecting Diabetic Retinopathy Based on Integrated Shallow Convolutional Neural Networks

Contact Info

Product

Resources

About