High-resolution retinal imaging system: diagnostic accuracy and usability

The development of high-resolution retinal imaging systems is critical for enhancing the diagnostic accuracy and usability of tools used in detecting glaucoma and managing other ophthalmic and systemic diseases. This study evaluates a novel high-resolution retinal imaging system by comparing its diagnostic performance in detecting glaucoma with AutoMorph, a leading retinal vessel segmentation tool with available online code for reproducibility. The system's diagnostic accuracy was assessed using Area Under the Curve (AUC) metrics, with our system (HRVIAS) achieving a superior AUC of 0.7048 compared to AutoMorph's AUC of 0.6560. Additionally, a usability study was conducted using the System Usability Scale (SUS), where participants rated the system highly, with the majority of scores clustering around 80 to 85, indicating strong user satisfaction. These findings demonstrate that the proposed system not only improves the diagnostic accuracy of detecting glaucoma but also offers a user-friendly interface, making it a valuable tool for clinical and research applications in retinal imaging.