AI-Driven Diabetic Retinopathy Diagnosis Enhancement through Image Processing and Salp Swarm Algorithm-Optimized Ensemble Network

TL;DR

This work tackles diabetic retinopathy classification by proposing an SSA-optimized ensemble that fuses multi-resolution retinal features through CLAHE, Gamma correction, and Discrete Wavelet Transform. It selects three strong backbones—DenseNet169, MobileNetV1, and Xception—augments them with an improved residual block, and optimizes their ensemble weights using Salp Swarm Algorithm to maximize accuracy on the Kaggle APTOS 2019 dataset, achieving 89.07% overall accuracy. The approach is validated with comprehensive metrics (ROC, PR, confusion matrix) and statistical testing (McNemar’s test), showing superiority over individual models and several existing methods, with additional validation on an external DDR dataset. The methodology offers a robust, interpretable pipeline that leverages advanced image processing and metaheuristic optimization to enhance DR detection, with future work targeting real-time mobile applications and further exploration of fusion techniques and class-imbalance mitigation.

Abstract

Diabetic retinopathy is a leading cause of blindness in diabetic patients and early detection plays a crucial role in preventing vision loss. Traditional diagnostic methods are often time-consuming and prone to errors. The emergence of deep learning techniques has provided innovative solutions to improve diagnostic efficiency. However, single deep learning models frequently face issues related to extracting key features from complex retinal images. To handle this problem, we present an effective ensemble method for DR diagnosis comprising four main phases: image pre-processing, selection of backbone pre-trained models, feature enhancement, and optimization. Our methodology initiates with the pre-processing phase, where we apply CLAHE to enhance image contrast and Gamma correction is then used to adjust the brightness for better feature recognition. We then apply Discrete Wavelet Transform (DWT) for image fusion by combining multi-resolution details to create a richer dataset. Then, we selected three pre-trained models with the best performance named DenseNet169, MobileNetV1, and Xception for diverse feature extraction. To further improve feature extraction, an improved residual block is integrated into each model. Finally, the predictions from these base models are then aggregated using weighted ensemble approach, with the weights optimized by using Salp Swarm Algorithm (SSA).SSA intelligently explores the weight space and finds the optimal configuration of base architectures to maximize the performance of the ensemble model. The proposed model is evaluated on the multiclass Kaggle APTOS 2019 dataset and obtained 88.52% accuracy.

Figures (8)

• Figure 1: Overview of image Fusion Framework
• Figure 2: Architecture overview of original (a) & improved block (b)
• Figure 3: Flowchart illustrating the step-by-step process of SSA
• Figure 4: Illustration of the Proposed Methodology for Diabetic Retinopathy Detection Using the SSA-Driven Ensemble approach
• Figure 5: Illustration of the data pre-processing and augmentation steps for model training