Search papers, labs, and topics across Lattice.
This paper introduces a lightweight deep learning framework based on MobileNetV3 with a Consistent Rank Logits (CORAL) head for automated diabetic retinopathy (DR) severity assessment from fundus images. The framework models the ordered progression of DR to improve diagnostic accuracy while maintaining computational efficiency for mobile deployment. Experiments on a combined APTOS 2019 and IDRiD dataset demonstrate a Quadratic Weighted Kappa (QWK) score of 0.9019 and an accuracy of 80.03%, with further optimizations including model calibration for real-world deployment.
Achieve near-expert-level diabetic retinopathy diagnosis on mobile devices with a MobileNetV3-based model, opening the door to scalable and accessible early screening.
Diabetic Retinopathy (DR) is a major cause of vision impairment worldwide. However, manual diagnosis is often time-consuming and prone to errors, leading to delays in screening. This paper presents a lightweight automated deep learning framework for efficient assessment of DR severity from digital fundus images. We use a MobileNetV3 architecture with a Consistent Rank Logits (CORAL) head to model the ordered progression of disease while maintaining computational efficiency for resource-constrained environments. The model is trained and validated on a combined dataset of APTOS 2019 and IDRiD images using a preprocessing pipeline including circular cropping and illumination normalization. Extensive experiments including 3-fold cross-validation and ablation studies demonstrate strong performance. The model achieves a Quadratic Weighted Kappa (QWK) score of 0.9019 and an accuracy of 80.03 percent. Additionally, we address real-world deployment challenges through model calibration to reduce overconfidence and optimization for mobile devices. The proposed system provides a scalable and practical tool for early-stage diabetic retinopathy screening.
