Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DeepEyeNet: Adaptive Genetic Bayesian Algorithm Based Hybrid ConvNeXtTiny Framework For Multi-Feature Glaucoma Eye Diagnosis

Angshuman Roy, Anuvab Sen, Soumyajit Gupta, Soham Haldar, Subhrajit Deb, Taraka Nithin Vankala, Arkapravo Das

TL;DR

This work tackles early glaucoma detection from fundus images, addressing variability in image quality and accessibility. It introduces DeepEyeNet, a unified framework that combines adaptive fundus image standardization, U-Net optic disc/cup segmentation, multi-feature extraction (anatomical, NR rim, texture, and vessel features), and a ConvNeXtTiny classifier optimized by Adaptive Genetic Bayesian Optimization. The approach achieves a notable $95.84\%$ accuracy and $0.9848$ AUC on EyePACS-AIROGS-light-V2, outperforming several state-of-the-art models and validating the effectiveness of AGBO for hyperparameter tuning. The framework holds promise for scalable, AI-assisted glaucoma screening in clinical and under-resourced settings, with potential extensions to other ocular diseases.

Abstract

Glaucoma is a leading cause of irreversible blindness worldwide, emphasizing the critical need for early detection and intervention. In this paper, we present DeepEyeNet, a novel and comprehensive framework for automated glaucoma detection using retinal fundus images. Our approach integrates advanced image standardization through dynamic thresholding, precise optic disc and cup segmentation via a U-Net model, and comprehensive feature extraction encompassing anatomical and texture-based features. We employ a customized ConvNeXtTiny based Convolutional Neural Network (CNN) classifier, optimized using our Adaptive Genetic Bayesian Optimization (AGBO) algorithm. This proposed AGBO algorithm balances exploration and exploitation in hyperparameter tuning, leading to significant performance improvements. Experimental results on the EyePACS-AIROGS-light-V2 dataset demonstrate that DeepEyeNet achieves a high classification accuracy of 95.84%, which was possible due to the effective optimization provided by the novel AGBO algorithm, outperforming existing methods. The integration of sophisticated image processing techniques, deep learning, and optimized hyperparameter tuning through our proposed AGBO algorithm positions DeepEyeNet as a promising tool for early glaucoma detection in clinical settings.

DeepEyeNet: Adaptive Genetic Bayesian Algorithm Based Hybrid ConvNeXtTiny Framework For Multi-Feature Glaucoma Eye Diagnosis

TL;DR

This work tackles early glaucoma detection from fundus images, addressing variability in image quality and accessibility. It introduces DeepEyeNet, a unified framework that combines adaptive fundus image standardization, U-Net optic disc/cup segmentation, multi-feature extraction (anatomical, NR rim, texture, and vessel features), and a ConvNeXtTiny classifier optimized by Adaptive Genetic Bayesian Optimization. The approach achieves a notable accuracy and AUC on EyePACS-AIROGS-light-V2, outperforming several state-of-the-art models and validating the effectiveness of AGBO for hyperparameter tuning. The framework holds promise for scalable, AI-assisted glaucoma screening in clinical and under-resourced settings, with potential extensions to other ocular diseases.

Abstract

Glaucoma is a leading cause of irreversible blindness worldwide, emphasizing the critical need for early detection and intervention. In this paper, we present DeepEyeNet, a novel and comprehensive framework for automated glaucoma detection using retinal fundus images. Our approach integrates advanced image standardization through dynamic thresholding, precise optic disc and cup segmentation via a U-Net model, and comprehensive feature extraction encompassing anatomical and texture-based features. We employ a customized ConvNeXtTiny based Convolutional Neural Network (CNN) classifier, optimized using our Adaptive Genetic Bayesian Optimization (AGBO) algorithm. This proposed AGBO algorithm balances exploration and exploitation in hyperparameter tuning, leading to significant performance improvements. Experimental results on the EyePACS-AIROGS-light-V2 dataset demonstrate that DeepEyeNet achieves a high classification accuracy of 95.84%, which was possible due to the effective optimization provided by the novel AGBO algorithm, outperforming existing methods. The integration of sophisticated image processing techniques, deep learning, and optimized hyperparameter tuning through our proposed AGBO algorithm positions DeepEyeNet as a promising tool for early glaucoma detection in clinical settings.
Paper Structure (27 sections, 4 equations, 8 figures, 3 tables, 2 algorithms)

This paper contains 27 sections, 4 equations, 8 figures, 3 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Proposed Methodology
  3. Dataset and Preprocessing
  4. Fundus Image Standardization
  5. Overall Framework
  6. Optic Disc and Cup Segmentation
  7. Feature Extraction
  8. Anatomical Features
  9. Neuroretinal Rim Metrics
  10. Texture Features
  11. Vessel Features
  12. Classification Model
  13. Network Architecture
  14. Model Details
  15. Custom Loss Function
  16. ...and 12 more sections

Figures (8)

  • Figure 1: Sample images from the EyePACS-AIROGS-light-V2 dataset showcasing RG and NRG.
  • Figure 2: Fundus images illustrating normal and glaucomatous eyes with emphasis on key optic structures.
  • Figure 3: Workflow of the proposed DeepEyeNet framework.
  • Figure 4: Architecture of ConvNeXtTiny-based classification network.
  • Figure 5: Flowchart of the proposed AGBO hyperparameter tuning process.
  • ...and 3 more figures