Portfolio

MetaVIn is a groundbreaking deep learning framework designed to mitigate atmospheric turbulence in long-range imaging by integrating meteorological data with visual information. This project, published in WACV 2025, represents a significant advancement in the field of atmospheric turbulence mitigati...

We designed and implemented a multi-stage pipeline that combines geometric consistency checks with region growing and mask refinement. Our approach starts with calculating bidirectional optical flow, which is then stabilized using temporal averaging to reduce turbulence-induced errors. We use epipol...

We developed a novel physics-based deep learning approach for estimating atmospheric turbulence strength from video data. Our method combines classical image gradient techniques with modern deep learning to accurately measure the refractive index structure constant (Cn2) that characterizes atmospher...

We developed a novel spectroscopic approach for detecting COVID-19 from saliva samples using absorption and emission spectroscopy. Our method analyzes spectral signatures to identify COVID-19 infection without requiring complex sample preparation or reagents, achieving high classification accuracy t...

Developed an architecture to analyze images of custom contact lenses and predict blood glucose level with 85% accuracy. Measured with spectroscopy system and got better results from image analysis based methods....

We developed a comprehensive analysis system to track and analyze the spread of COVID-19 related information across social media platforms. Our team processed 387,000 tweets, analyzed 25 news portals, and tracked 35 trusted health officials to understand how misinformation impacts public sentiment d...

During my MS at Gwangju Institute of Science and Technology, I worked in the Neurophotonics Lab under Professor Euiheon Chung's supervision. The lab specialized in developing advanced optical imaging systems for microscopic visualization. I contributed to multiple optical microscopy projects, focusi...

I developed a deep learning-based automated system for analyzing meibomian gland (MG) dysfunction from infrared meibography images. The system performs MG segmentation, eyelid segmentation, quantitative analysis of MG morphology, and meiboscore prediction. Additionally, it removes specular reflectio...

We developed a novel neural network architecture combining CNN and LSTM in parallel for motor imagery brain-computer interface (MI-BCI) classification. Our approach achieved 87.68% accuracy and significantly outperformed existing methods by integrating spatial, temporal, and middle-layer features in...

I developed a novel segment-then-restore pipeline called Turb-Seg-Res for restoring videos degraded by atmospheric turbulence, particularly focusing on dynamic scenes. The system first segments moving objects from static backgrounds and then applies targeted restoration techniques to each component....