• Journal of Biomedical Engineering Research
• /
• v.40 no.2
• /
• pp.48-54
• /
• 2019

This paper proposed a method of validation data augmentation for improving the grading accuracy of diabetic macular edema (DME) using deep learning. The data augmentation technique is basically applied in order to secure diversity of data by transforming one image to several images through random translation, rotation, scaling and reflection in preparation of input data of the deep neural network (DNN). In this paper, we apply this technique in the validation process of the trained DNN, and improve the grading accuracy by combining the classification results of the augmented images. To verify the effectiveness, 1,200 retinal images of Messidor dataset was divided into training and validation data at the ratio 7:3. By applying random augmentation to 359 validation data, $1.61{\pm}0.55%$ accuracy improvement was achieved in the case of six times augmentation (N=6). This simple method has shown that the accuracy can be improved in the N range from 2 to 6 with the correlation coefficient of 0.5667. Therefore, it is expected to help improve the diagnostic accuracy of DME with the grading information provided by the proposed DNN.

• Vacuum Magazine
• /
• v.4 no.3
• /
• pp.12-15
• /
• 2017

Surface enhanced Raman scattering (SERS) is considered as one of promising medical and diagnostic technologies. The SERS effect is caused by the localized surface plasmon resonance (LSPR) from metal nanoparticles with narrow hot spots. The mechanism of LSPR, development of nanostructure fabrication, and corresponding researches are discussed. The flexible, label-free, low-cost, and highly-sensitive Au/ZnONRs/G is introduced. The Au/ZnONRs/G detects and distinguishes cataract, age-related macular degeneration, and diabetic macular edema from aqueous humor. Comprehension of SERS provides further improvement in bio sensing technology including early diagnosis and prolonged life expectancy.realize highly stretchable electrodes.

• Journal of The Korean Ophthalmological Society
• /
• v.60 no.2
• /
• pp.144-151
• /
• 2019

Purpose: Intravitreal aflibercept, ranibizumab, bevacizumab, and dexamethasone are the most widely used drugs in the treatment of diabetic macular edema (DME). The aim of this study was to compare the efficacy and safety of anti-vascular endothelial growth factors and dexamethasone for the treatment of DME. Methods: There were nine previous systematic reviews on this topic; we updated these high-quality reviews. Seven studies were added to two studies following a literature search. Efficacy outcomes were 1) average improvement in visual acuity, 2) proportion of patients who experienced an improvement in vision (an increase in best-corrected visual acuity (BCVA) of ${\geq}15$ in the Early Treatment Diabetic Retinopathy Study [ETDRS]), and 3) proportion of patients who experienced worsening vision (a decrease in BCVA of ${\geq}15$ in the ETDRS). Safety outcomes included systemic adverse events and ocular-related adverse events. Results: The mean difference in the BCVA for ranibizumab versus bevacizumab treatment was 0.16 (95% confidence interval [CI]: -0.02, 0.34), and that for ranibizumab versus aflibercept was -0.08 (95% CI: -0.26, 0.10). The mean difference in the change of BCVA for aflibercept versus ranibizumab was -0.20 (95% CI: -0.40, -0.01), and that for aflibercept versus bevacizumab was -0.34 (95% CI: -0.53, -0.14). Other efficacy outcomes showed similar trends, and there was no significant difference between treatments. There was also no significant difference in both systemic and ocular adverse events rates between the treatments. Conclusions: In DME patients, the efficacy of aflibercept was found to be higher with respect to BCVA changes compared with ranibizumab or bevacizumab. However, there were no significant difference in terms of visual acuity improvement or visual acuity of more than 15 letters, nor in terms of anti-vascular endothelial growth factors (as a safety outcome).