Park Sang Min

Explaining the Rationale of Deep Learning Glaucoma Decisions with Adversarial Examples.

Adversarial explanation increased the explainability over GradCAM, a conventional heatmap-based explanation method. Adversarial explanation may help medical professionals understand more clearly the rationale of DLMs when using them for clinical decisions.

Reproducibility of Retinal Nerve Fiber Layer and Macular Ganglion Cell Layer Thickness Measurements by Optical Coherence Tomography in Myopic Eyes.

Long-term reproducibility of RNFL and GCIPL thickness measurements in moderate-to-high myopia was comparable to that of low-to-no myopia, except RNFL thickness in the temporal quadrant.