Using MobileNetV2 to Classify the Severity of Diabetic Retinopathy

Sheikh, Sarah; Qidwai, Uvais

doi:10.5013/ijssst.a.21.02.16

Cited by 14 publications

(6 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We have tested the model using different hyper-parameters to improve the performance. Some studies in the literature such as [22] has gained an accuracy value of 91.68% which is higher than the proposed model. In [22], the authors used custom data that have been selected in good quality to contribute positively in the training process.…”

Section: Resultsmentioning

confidence: 71%

See 1 more Smart Citation

Fundus Images Classification of Diabetic Retinopathy using MobileNetV2

Yassin¹

2023

IJCSMC

View full text Add to dashboard Cite

Diabetic retinopathy (DR) is a chronic eye disease and the main reason of blindness between adults. Manual DR screening methods require skilled readers, effort, and time. Automated DR detection and classification of its severity is important for early and effective disease control. Mobile devices can facilitate regular screening of retinal fundus images using automated lightweight architectures techniques. While, many automated techniques have been proposed in the literature, new models are needed to improve the accuracy and to facilitate the accessibility to such models. In this study, MobileNetV2 is used to classify five classes of DR severity that includes no DR, mild DR, moderate DR, severe DR, and proliferative DR. Colour fundus images are enhanced using contrast-limited adaptive histogram equalization (CLAHE) to defeat bad lighting conditions. MobileNetV2 is used to extract image features. A new classification head based on batch normalization and L2 regularization is applied on the extracted features. The model is tested using augmented data, non-augmented data, fine-tuning, and without fine-tuning. The experiments indicate the importance of augmentation and fine-tuning for improving the evaluation metrics. The experimental results show the superiority of the proposed technique compared to recent techniques in literature with training, validation, and test accuracies of 98.65, 87.94%, and 87.67% respectively.

show abstract

Section: Resultsmentioning

confidence: 71%

“…They classified DR into 2, 3, and 5 classes with achieved accuracies of 61.3%, 60.3% and 37.7% respectively. Sheikh and Qidwai [22] proposed a DR classification system based on MobileNetV2 architecture. They used a custom dataset to classify DR severities into five classes, where good quality images were selected for training process.…”

Section: IImentioning

confidence: 99%

Fundus Images Classification of Diabetic Retinopathy using MobileNetV2

Yassin¹

2023

IJCSMC

View full text Add to dashboard Cite

show abstract

“…Gao Jinfeng et al 14 two deep CNN models used an ensemble method to detect all the DR stages by using balanced and unbalanced datasets. The outcome demonstrated that models outperform more sophisticated techniques like the Kaggle datasets in contrast to how well they now detect all stages of DR. By implementing a light-weight mobile network and assessing the effectiveness of their classifier, Sheikh et al 15 employed a novel approach. MobileNetV2 was constructed as a light-weight, mobile-friendly architecture and trained using datasets of diabetic retinal fundus images.…”

Section: Related Workmentioning

confidence: 99%

Deploying efficient net batch normalizations (BNs) for grading diabetic retinopathy severity levels from fundus images

Batool,

Gilani,

Waris

et al. 2023

Sci Rep

View full text Add to dashboard Cite

Diabetic retinopathy (DR) is one of the main causes of blindness in people around the world. Early diagnosis and treatment of DR can be accomplished by organizing large regular screening programs. Still, it is difficult to spot diabetic retinopathy timely because the situation might not indicate signs in the primary stages of the disease. Due to a drastic increase in diabetic patients, there is an urgent need for efficient diabetic retinopathy detecting systems. Auto-encoders, sparse coding, and limited Boltzmann machines were used as a few past deep learning (DL) techniques and features for the classification of DR. Convolutional Neural Networks (CNN) have been identified as a promising solution for detecting and classifying DR. We employ the deep learning capabilities of efficient net batch normalization (BNs) pre-trained models to automatically acquire discriminative features from fundus images. However, we successfully achieved F1 scores above 80% on all efficient net BNs in the EYE-PACS dataset (calculated F1 score for DeepDRiD another dataset) and the results are better than previous studies. In this paper, we improved the accuracy and F1 score of the efficient net BNs pre-trained models on the EYE-PACS dataset by applying a Gaussian Smooth filter and data augmentation transforms. Using our proposed technique, we have achieved F1 scores of 84% and 87% for EYE-PACS and DeepDRiD.

show abstract

“…Qin and Li [17] introduced a novel technique for detecting the facemask-wearing nation called super resolution and classification network (SRCNet) to acquire an accuracy rate of 98.70%. Sheikh and Qidwai [18] proposed a new technique using a light mobile system and evaluated the execution of their classifier developed using MobileNetV2 with a remarkable degree of accuracy of 91.68%. Sethi et al [19] proposed an approach that integrated a single-stage and two-stage detector.…”

Section: Introductionmentioning

confidence: 99%

Face mask detection using deep learning on NVIDIA Jetson Nano

Hassan

Abed²,

Abdalla³

2022

IJECE

View full text Add to dashboard Cite

<span>In December 2019, the coronavirus pandemic started. Coronavirus desease-19 (COVID-19) is transmitted directly from contaminated surfaces via direct touch. To combat the virus, a multitude of equipment is needed. Masks are a vital element of personal protection in crowded places. As a result, determining if a person is wearing a face mask is critical to assimilating to contemporary society. To accomplish the objective, the model presented in this paper used deep learning libraries and OpenCV. This approach was chosen for safety concerns due to its high resource efficiency during deployment. The classifier was built using the MobileNetV2 structure, which was designed to be lightweight and capable of being utilized in embedded devices such as the NVIDIA Jetson Nano to do real-time mask recognition. The stages of model construction were collecting, pre-processing, splitting data, creating the model, training the model, and applying the model. This system utilized image processing techniques and deep learning to process a live video feed. When someone is not wearing a mask, the output eventually produces an alarm sound through a built-in buzzer. Experimental results and testing were used to verify the suggested system's performance. Including both training and testing, the achieved recognition rate was 99%.</span>

show abstract

Using MobileNetV2 to Classify the Severity of Diabetic Retinopathy

Cited by 14 publications

References 17 publications

Fundus Images Classification of Diabetic Retinopathy using MobileNetV2

Fundus Images Classification of Diabetic Retinopathy using MobileNetV2

Deploying efficient net batch normalizations (BNs) for grading diabetic retinopathy severity levels from fundus images

Face mask detection using deep learning on NVIDIA Jetson Nano

Contact Info

Product

Resources

About