Development of preprocessing methods and revised <scp>EfficientNet</scp> for diabetic retinopathy detection

Lin, Chun‐Ling; Jiang, Zhi‐Xiang

doi:10.1002/ima.22874

Cited by 3 publications

(2 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The paper published in 2023 represents our team's research on DR classification [6]. It emphasized the impact of varying image quality due to different conditions on the efficacy of training models to classify DR stages, spanning from 0-No DR to 4-Proliferative DR. A proposed preprocessing method enhanced image features, effectively expanding the training dataset.…”

Section: Our Previous Workmentioning

confidence: 99%

“…This study proceeds to a comparison with prior research that utilized the same ATPOS dataset for training. The paper published in 2023 represents our team's research on DR classification [6], in which we employed preprocessing quality filters to eliminate low-quality images and trained an EfficientNet model. S. H. Kassani et al [5] employed minimal pooling preprocessing on retinal images and employed a variety of pretrained models for training.…”

Section: Comparison With Prior Researchmentioning

confidence: 99%

See 1 more Smart Citation

Enhancing Diabetic Retinopathy Detection Using Pixel Color Amplification and EfficientNetV2: A Novel Approach for Early Disease Identification

Kao,

Lin

2024

Electronics

Self Cite

View full text Add to dashboard Cite

Diabetic retinopathy (DR) is a severe complication of diabetes, causing damage to retinal blood vessels due to high blood sugar levels. Early detection is crucial but often requires significant time and expertise from ophthalmologists. While artificial intelligence (AI) and image recognition hold promise for DR detection, inconsistent image quality poses a challenge. Our study presents a novel technique that integrates pixel color amplification and EfficientNetV2 to enhance fundus image attributes, aiming to address issues related to image quality and achieving superior performance in DR detection. Leveraging EfficientNetV2, an advanced convolutional neural network (CNN) architecture, we achieve 84% multiclass accuracy and 99% binary accuracy, surpassing various other CNN models, including VGG16-fc1, VGG16-fc2, NASNet, Xception, Inception ResNetV2, EfficientNet, InceptionV3, MobileNet, and ResNet50. Our research tackles the critical challenge of early detection of DR, essential for preventing vision loss. This advancement holds the potential to enhance the efficiency and accuracy of DR classification, potentially alleviating the burden on medical professionals and ultimately improving the quality of life for individuals at risk of vision loss.

show abstract

Section: Our Previous Workmentioning

confidence: 99%

Section: Comparison With Prior Researchmentioning

confidence: 99%

Enhancing Diabetic Retinopathy Detection Using Pixel Color Amplification and EfficientNetV2: A Novel Approach for Early Disease Identification

Kao,

Lin

2024

Electronics

Self Cite

View full text Add to dashboard Cite

show abstract

An explainable transfer learning framework for multi-classification of lung diseases in chest X-rays

Patel,

Murugan,

Srivastava

et al. 2024

Alexandria Engineering Journal

View full text Add to dashboard Cite

Segmentation Using the IC2T Model and Classification of Diabetic Retinopathy Using the Rock Hyrax Swarm-Based Coordination Attention Mechanism

Jagadesh,

Karthik,

Siri

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Diabetic Retinopathy (DR) evaluations are increasingly being automated using artificial intelligence. Diabetes-related retinal vascular disease is a major cause of blindness and visual impairment worldwide. Therefore, automated DR detection devices would greatly aid in reducing visual impairment due to DR through early screening and treatment. Researchers have provided many techniques for picking out abnormalities in retinal images during the past several years. In the past, automated methods for diagnosing diabetic retinopathy required a human to extract information from retinal images before passing them on to a classifier. This study takes a novel two-pronged approach to automated DR classification to solve the issues. Due to the low positive instance percentage of existing asymmetric, we segment O.D.s and B.V.s with an enhanced version of an improved contoured convolutional transformer (IC2T). We develop a contoured optical disc (OD), a blood vessels (BV) detection module, and a dual convolutional transformer block that combines local and global contexts to make trustworthy associations. A second-stage Improved Coordination Attention Mechanism (ICAM) network is trained to recognize retinal biomarkers for DR such as microaneurysms (M.A.), haemorrhages (H.M.), and exudates (EX). With an average accuracy of 96%, 97%, and 98% on EyePACS-1, Messidor-2, and DIARETDB0, respectively, the suggested technique has proven itself to be at the field's cutting edge. Comprehensive testing and comparisons to established methods support the proposed strategy.

show abstract

Development of preprocessing methods and revised EfficientNet for diabetic retinopathy detection

Cited by 3 publications

References 47 publications

Enhancing Diabetic Retinopathy Detection Using Pixel Color Amplification and EfficientNetV2: A Novel Approach for Early Disease Identification

Enhancing Diabetic Retinopathy Detection Using Pixel Color Amplification and EfficientNetV2: A Novel Approach for Early Disease Identification

An explainable transfer learning framework for multi-classification of lung diseases in chest X-rays

Segmentation Using the IC2T Model and Classification of Diabetic Retinopathy Using the Rock Hyrax Swarm-Based Coordination Attention Mechanism

Contact Info

Product

Resources

About