👋Daehee Park
Ph.D. student
About Me
Daehee Park is a Ph.D. candidate at Korea Advanced Institute of Science and Technology (KAIST), advised by Prof. Kuk-Jin Yoon. He received B.S. and M.S. degree from KAIST in 2018 and 2020. He had research intern at Naver Labs (2021) and Qualcomm (2024). His research interests include expanding applicability of deep learning and computer vision for autonomous driving. His recent researchs are focused on modeling movement of agents under chellenging real-world problems (e.g. complex interaction, domain shift, and long-tail problem)
Recent News
Experience
Qualcomm
Deep Learning R&D Intern- Work with the US team in remote
- Development of deep learning models to predict trajectories and/or intentions for road users
Naver Labs
Research Intern- Development of a deep learning model that predict future trajectory of road agents for autonomous driving
- Development of realistic driving simulator using deep trajectory prediction network
Education
PhD in Mechanical Engineering
KAISTMS in Mechanical Engineering
KAISTBSc in Mechanical Engineering and Business and Technology Management
KAIST
Featured Publications
Recent Publications
(2024).
Diffusion-Guided Weakly Supervised Semantic Segmentation.
ECCV 2024.
(2024).
Multi-agent Long-term 3D Human Pose Forecasting via Interaction-aware Trajectory Conditioning.
CVPR 2024 (Hightlight).
(2024).
T4P: Test-Time Training of Trajectory Prediction via Masked Autoencoder and Actor-specific Token Memory.
CVPR 2024.
(2024).
Improving Transferability for Cross-domain Trajectory Prediction via Neural Stochastic Differential Equation.
AAAI 2024.
(2023).
Leveraging Future Relationship Reasoning for Vehicle Trajectory Prediction.
ICLR 2023.
(2022).
BIPS: Bi-modal Indoor Panorama Synthesis via Residual Depth-Aided Adversarial Learning.
ECCV 2022.
(2021).
Unlocking the Potential of Ordinary Classifier: Class-Specific Adversarial Erasing Framework for Weakly Supervised Semantic Segmentation.
ICCV 2021.
(2020).
Identifying Reflected Images From Object Detector in Indoor Environment Utilizing Depth Information.
IEEE Robotics and Automation Letters and ICRA 2021.