Minsik Jeon

I am a Robotics & Computer Vision researcher at the Agency for Defense Development (ADD) in the South Korea. My research interests lie in computer vision and robotics, with a special focus on robot perception that is generalizable and adaptable under diverse environments. I completed my bachelor's degree (Summa Cum Laude) in School of Computing at KAIST (Korea Advanced Institute of Science and Technology).

Email  /  CV  /  Scholar  /  LinkedIn

Research Interests

I am interested in the intersection of computer vision and robotics, with the goal of developing robust perception algorithms for robots across diverse environments. I am particularly interested in methods to learn generalizable features from unlabeled and multimodal datasets, as well as adaptation methods to adjust the models to new target domains.
My current & future interests are as follows:

● Domain Adaptation & Generalization
● Multi-modal, Open-world perception
● Self-supervised Learning
● Uncertainty-aware perception
● Off-road applications in robotics

Education

Publications

A method for Open-World object detection that improves novel object detection and semantically rich feature extraction by leveraging Vision Foundation Models.

An UDA framework for object detection that can effectively adapt to real-world adverse weather conditions by addressing Style & Weather gaps separately.

Projects

- Developing perception and control algorithms for off-road navigation: Robust moving object detection & tracking in off-road environment, semantic terrain classification map through LiDAR-Camera integration.
- Examined path planning and control algorithms for navigation in complex environments.

- Develop a BEV traversability map by combining traversability estimates from UAV and multiple UGVs for off-road autonomous driving, including sensor data acquisition and integration, UAV image registration, and uncertainty-aware mapping.
- Building a generalizable LiDAR semantic segmentation model across various LiDAR sensor configurations on robotic platforms.

- Developed an LiDAR-IR fusion method for robust object detection under visibility-constrainted scenarios.

- Developed an open-set 2D & 3D object detection method for LiDAR pointcloud.

Participated in the Indy Autonomous Challenge (IAC), the first autonomous car racing competition to be held in Indianapolis, as an intern for Team KAIST. As part of a perception team, I tested algorithms using camera and lidar in a simulation environment.

Developed a removal network that addresses the line artifact of SCAPE microscopy image, using CycleGAN-like model and transferring the image into a CFM style.

In Performance analysis team, configured On-board RAID using SSDs and conducted performance evaluations for each RAID option.

Template from Jon Barron's wonderful work.