Minseok Kim - Homepage

Minseok Kim (김민석, 金珉奭)

Hello, I am a Master student at UNIST, where I research on 3D computer vision. During my master's degree, I worked at the Perceiving Systems and OSLab of Max Planck Institute for Intelligent Systems.

Before that, I received my Bachelor's degree from Chung-Ang University. During the Bachelor, I was selected for the K-SW Square Global Education Program. So I conducted a team project under the theme of reinforcement learning at Purdue University.

Email / LinkedIn / Github / Google Scholar

Research

My research interests lie in the fields of 3D computer vision. In particular, I focus on realistic movements using humans and animals. Furthermore, I'm interested in the 3D representation and non-rigid motion estimation/generation.

Pose-Guided 3D Human Generation in Indoor Scene
Minseok Kim, Changwoo Kang, Jeongin Park, Kyungdon Joo
Association for the Advancement of Artificial Intelligence (AAAI), 2023
paper / project page / video

Address the problem of scene-aware 3D human generation based on interactions.
Pay attention to the fact that physical contact between 3D human and scene.
Propose normal alignment between human and scene.

Selected Projects

	3D Scene Graph Generation using 360° Panoramic Images Minseok Kim 3D Vision Course Project, 2021 Propose a 3D scene graph generation method using RGB-D panoramic image. Apply Mask-RCNN on the panoramic image for identifying objects. Define a 3D relationship between two objects using cosine similarity.
	Image Inpainting using Edge and Panoptic Features Minseok Kim, Jeongin Park, Solang Kim Deep Generative Model Course Project, 2021 Propose the image inpainting method using edge and panoptic information. Extract the edge and panoptic features given RGB images.
	Motion Planning for Starcraft Game Units using Reinforcement Learning Minseok Kim, Jeongwoo Lee, Minseop Kim, Hankyeol Kyung, Hanam Hwang, Jeongyeon Lee, Minjeong Jung AIC & SAMSUNG COMPANY Project, 2019/2020 Propose the motion planning algorithm using reinforcement learning. Implement the single-agent, multi-agent algorithms in an obstacle dense game environment.
	Motion Planning for Online Soccer Game Units using Reinforcement Learning Minseok Kim, Junseon Kim, Seulgi Choi, Inseok Ryu IITP-Purdue University Summer Project, 2019 Propose the motion planning algorithm using online soccer game. Design the game strategy and implement it using reinforcement learning algorithm.

Teaching

Software Engineering Tutor, Chung-Ang University, 2019/2020

template adapted from this awesome website
Last updated: Apr 2023