A Hybrid GAN-BiGRU Model Enhanced by African Buffalo Optimization for Diabetic Retinopathy Detection DOI Open Access

P Sasikala,

Sushil Dohare,

Mohammed Saleh Al Ansari

et al.

International Journal of Advanced Computer Science and Applications, Journal Year: 2024, Volume and Issue: 15(1)

Published: Jan. 1, 2024

Diabetic retinopathy (DR) is a severe complication of diabetes mellitus, leading to vision impairment or even blindness if not diagnosed and treated early. A manual inspection the patient's retina conventional way for diagnosing diabetic retinopathy. This study offers novel method identification in medical diagnosis. Using hybrid Generative Adversarial Network (GAN) Bidirectional Gated Recurrent Unit (BiGRU) model, further refined using African Buffalo Optimization algorithm, model's capacity identify minute patterns suggestive improved by GAN's skill extracting complex characteristics from retinal pictures. The technique feature extraction plays critical role revealing information that may be hidden yet essential precise Then, BiGRU part works on have been extracted, efficiently maintaining temporal relationships, enabling thorough absorption. combination capabilities with BiGRU's sequential processing capability creates synergistic interaction gives model comprehensive grasp Moreover, utilized optimize performance accuracy fine-tuning its parameters. current study, which uses Python, obtains 98.5% rate demonstrates amazing ability reach high levels Retinopathy Detection.

Language: Английский

Deep learning for retinal vessel segmentation: a systematic review of techniques and applications DOI
Zhihui Liu, Mohd Shahrizal Sunar, Tan Tian Swee

et al.

Medical & Biological Engineering & Computing, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 18, 2025

Language: Английский

Citations

0
Applications of generative adversarial networks in the diagnosis, prognosis, and treatment of ophthalmic diseases DOI Creative Commons
Robert Doorly, Joshua Ong, Ethan Waisberg

et al.

Graefe s Archive for Clinical and Experimental Ophthalmology, Journal Year: 2025, Volume and Issue: unknown

Published: April 22, 2025

Abstract Purpose Generative adversarial networks (GANs) are key components of many artificial intelligence (AI) systems that applied to image-informed bioengineering and medicine. GANs combat limitations facing deep learning models: small, unbalanced datasets containing few images severe disease. The predictive capacity conditional may also be extremely useful in managing disease on an individual basis. This narrative review focusses the application ophthalmology, order provide a critical account current state ongoing challenges for healthcare professionals allied scientists who interested this rapidly evolving field. Methods We performed search studies apply generative diagnosis, therapy prognosis eight eye diseases. These disparate tasks were selected highlight developments GAN techniques, differences common features aid practitioners future adopters field ophthalmology. Results we identified show have demonstrated to: generate realistic synthetic images, convert image modality, improve quality, enhance extraction relevant features, prognostic predictions based input other data. Conclusion broad range architectures considered describe how technology is meet different (including segmentation multi-modal imaging) particular relevance wide availability now facilitates entry new researchers However mainstream adoption clinical use remains contingent larger public widespread validation necessary regulatory oversight.

Language: Английский

Citations

0
Revolutionizing diabetic retinopathy diagnosis through advanced deep learning techniques: Harnessing the power of GAN model with transfer learning and the DiaGAN-CNN model DOI
Mohamed R. Shoaib, Heba M. Emara, Ahmed S. Mubarak

et al.

Biomedical Signal Processing and Control, Journal Year: 2024, Volume and Issue: 99, P. 106790 - 106790

Published: Sept. 12, 2024

Language: Английский

Citations

2
Diabetic Retinopathy Lesion Segmentation Method Based on Multi-Scale Attention and Lesion Perception DOI Creative Commons

Ye Bian,

Chengyong Si, Lei Wang

et al.

Algorithms, Journal Year: 2024, Volume and Issue: 17(4), P. 164 - 164

Published: April 19, 2024

The early diagnosis of diabetic retinopathy (DR) can effectively prevent irreversible vision loss and assist ophthalmologists in providing timely accurate treatment plans. However, the existing methods based on deep learning have a weak perception ability different scale information retinal fundus images, segmentation capability subtle lesions is also insufficient. This paper aims to address these issues proposes MLNet for DR lesion segmentation, which mainly consists Multi-Scale Attention Block (MSAB) Lesion Perception (LPB). MSAB designed capture multi-scale features while LPB perceives depth. In addition, novel function with tailored weight reduce influence imbalanced datasets algorithm. performance comparison between other state-of-the-art carried out DDR dataset DIARETDB1 dataset, achieves best results 51.81% mAUPR, 49.85% mDice, 37.19% mIoU 67.16% mAUPR 61.82% mDice dataset. generalization experiment IDRiD 59.54% among methods. show that has outstanding ability.

Language: Английский

Citations

1
A Hybrid GAN-BiGRU Model Enhanced by African Buffalo Optimization for Diabetic Retinopathy Detection DOI Open Access

P Sasikala,

Sushil Dohare,

Mohammed Saleh Al Ansari

et al.

International Journal of Advanced Computer Science and Applications, Journal Year: 2024, Volume and Issue: 15(1)

Published: Jan. 1, 2024

Diabetic retinopathy (DR) is a severe complication of diabetes mellitus, leading to vision impairment or even blindness if not diagnosed and treated early. A manual inspection the patient's retina conventional way for diagnosing diabetic retinopathy. This study offers novel method identification in medical diagnosis. Using hybrid Generative Adversarial Network (GAN) Bidirectional Gated Recurrent Unit (BiGRU) model, further refined using African Buffalo Optimization algorithm, model's capacity identify minute patterns suggestive improved by GAN's skill extracting complex characteristics from retinal pictures. The technique feature extraction plays critical role revealing information that may be hidden yet essential precise Then, BiGRU part works on have been extracted, efficiently maintaining temporal relationships, enabling thorough absorption. combination capabilities with BiGRU's sequential processing capability creates synergistic interaction gives model comprehensive grasp Moreover, utilized optimize performance accuracy fine-tuning its parameters. current study, which uses Python, obtains 98.5% rate demonstrates amazing ability reach high levels Retinopathy Detection.

Language: Английский

Citations

0