Involved in multiple research projects including D2USt3R, Cross-View Completion Models are Zero-shot Correspondence Estimators, Local All-Pair Correspondence for Point Tracking, and RAIN-GS: Relaxing Accurate Initialization Constraint for 3D Gaussian Splatting.
Education
KAIST, Advisor: Prof. Seungryong Kim
Background
1st-year integrated MS/PH.D student at KAIST under the supervision of Prof. Seungryong Kim. Research interests lie in computer vision and deep learning, particularly in depth estimation and NeRF.