About Me

I am currently a PhD student at Jeonbuk National University, advised by Prof. Sang Jun Lee. My research focuses on developing vision systems that address real-world challenges in unsupervised learning, depth estimation, knowledge distillation, uncertainty estimation and perception algorithms like Simultaneous Localization And Mapping (SLAM) for Advanced Driver Assistance System (ADAS) and autonomous driving.

I’m currently researching depth-estimation–based outdoor SLAM aimed at operating without LiDAR, with a natural extension to LiDAR-assisted pipelines (e.g., depth completion) for added robustness. 

My long-term goal is to enable deep learning models to internalize geometric principles and infer scene-level information (e.g., 3D structure, pose, uncertainty), then deploy these insights as trustworthy, robust algorithms in real-world settings. 

RGB-Based Visual–Inertial Odometry via Knowledge Distillation from Self-Supervised Depth Estimation with Foundation Models

Sensors, 2025 (IF: 3.5)

Jimin Song and Sang Jun Lee*

[Paper]

Uncertainty-Aware Depth Network for Visual Inertial Odometry of Mobile Robots

Sensors, 2024 (IF: 3.5)

Jimin Song, HyungGi Jo, Yongsik Jin and Sang Jun Lee*

[Paper]

Knowledge Distillation of Multi-scale Dense Prediction Transformer for Self-supervised Depth Estimation

Scientific Reports, 2023 (IF: 3.8)

Jimin Song and Sang Jun Lee*

[Paper] • [Code]

Visual Simultaneous Localization and Mapping using Filtered Depth Maps based on a Foundation Model

International Conference on Intelligent Robots and Systems (IROS) workshop, 2024

Jimin Song, HyungGi Jo, Yongsik Jin and Sang Jun Lee*

[Paper]