Physical AI

← Research Area

Physical AI

Deep Learning

  • Deep learning-based object recognition
  • Event-to-Edge Gaussian Splatting for pose-free 3D scene reconstruction
  • Fast and robust local feature matching using Mamba architecture
  • Cross-view pose optimization network inspired by PID controllers
  • Confusion-and-confidence-aware mixture model for context optimization
  • Contextual contrastive learning for pixel-level scene understanding
  • Mesh-refined unsupervised depth completion leveraging structural regularities from Visual SLAM, Multi-stage depth prediction
  • Range-only network for distance-based localization
  • Bio-inspired neural network
  • Spiking neural network

Related papers

  • Yunsoo Kim, Changki Sung, Dasol Hong, and Hyun Myung†, “E2EGS: Event-to-Edge Gaussian Splatting for Pose-Free 3D Reconstruction," in Proc. The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2026), 2026. (Accepted)
  • Juhye Park, Wooju Lee, Dasol Hong, Changki Sung, Youngwoo Seo, DongWan Kang, and Hyun Myung†, “VIRD: View-Invariant Representation through Dual-Axis Transformation for Cross-View Pose Estimation," in Proc. The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2026), 2026. (Accepted)
  • Kihwan Ryoo, Hyungtae Lim, and Hyun Myung†, “MambaGlue: Fast and Robust Local Feature Matching With Mamba," in Proc. IEEE Int’l Conf. on Robotics and Automation (ICRA), pp. 5758-5765, Atlanta, USA, May 2025. [Link]
  • Wooju Lee, Juhye Park, Dasol Hong, Changki Sung, Youngwoo Seo, DongWan Kang, and Hyun Myung†, “PIDLoc: Cross-View Pose Optimization Network Inspired by PID Controllers," in Proc. The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2025), Nashville, USA, Jun. 2025. [Link]
  • Dasol Hong, Wooju Lee, and Hyun Myung†, “CoCoA-Mix: Confusion-and-Confidence-Aware Mixture Model for Context Optimization," in Proc. International Conference on Machine Learning (ICML 2025), Vancouver, Canada, Jul. 2025. [Link]
  • Changki Sung, Wanhee Kim, Jungho An, WooJu Lee, Hyungtae Lim, and Hyun Myung†, “Contextrast: Contextual Contrastive Learning for Semantic Segmentation," in Proc. The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2024), Seattle, USA, Jun. 2024. [Link]
  • Jinwoo Jeon, Hyunjun Lim, Dong-Uk Seo, and Hyun Myung†, “Struct-MDC: Mesh-Refined Unsupervised Depth Completion Leveraging Structural Regularities from Visual SLAM," in Proc. 2022 IEEE/RSJ Int’l Conf. on Intelligent Robots and Systems (IROS), pp. 6391-6398, Kyoto, Japan, Oct. 2022. [Link]
  • Hyungtae Lim, Hyeonjae Gil, and Hyun Myung†, “MSDPN: Monocular Depth Prediction with Partial Laser Observation using Multi-stage Neural Networks," in Proc. IEEE/RSJ Int’l Conf. on Intelligent Robots and Systems (IROS), Las Vegas, USA (virtual), Oct. 2020. [Link]
  • Hyungtae Lim, Changgue Park, and Hyun Myung†, “RONet: Real-time Range-only Indoor Localization via Stacked Bidirectional LSTM with Residual Attention," in Proc. IEEE/RSJ Int’l Conf. on Intelligent Robots and Systems (IROS), pp.3241-3247, Macau, China, Nov. 2019. [Link]

Photos and Videos

Gesture recognition using only depth information

  • Machine learning-based and Deep learning-based human pose estimation (skeleton extraction) and gesture recognition.

Related papers

    • Seunghee Lee, Jungmo Koo, Hyungjin Kim, Kwangyik Jung, and Hyun Myung†, “A Robust Estimation of 2D Human Upper-body Poses using Fully Convolutional Network," in Proc. Int'l Conf. on Robot Intelligence Technology and Applications (RiTA), Daejeon, Korea, Dec. 2017. [Link]
    • Hangeun Kim, Sangwon Lee, Youngjae Kim, Serin Lee, Dongsung Lee, Jinsun Ju, and Hyun Myung†, “Weighted Joint-Based Human Behavior Recognition Algorithm using Only Depth Information for Low-Cost Intelligent Video-Surveillance System," Expert Systems With Applications, vol.45, pp.131-141, Sep. 2015. [Link]
    • Hanguen Kim, Sangwon Lee, Dongsung Lee, Soonmin Choi, Jinsun Ju, and Hyun Myung†, “Real-time human pose estimation and gesture recognition from depth images using superpixels and SVM classifier," Sensors, vol.15, no.6, pp.12410-12427, May 2015. [Link]

Photos and Videos

Reinforcement Learning for Robotics

  • Quadrupedal blind locomotion via Deep RL
  • Nominal controller reinforcement for tilting-rotor drones
  • Legged manipulator control for articulated objects
  • Fault-aware locomotion in rough terrains

Related papers

  • Seonghyeon Lim, Seunghyun Lee, Hyeonwoo Lee, I Made Aswin Nahrendra, Hyun Myung, “OpenHEART: Opening Heterogeneous Articulated Objects with a Legged Manipulator," in Proc. IEEE Int'l Conf. on Robotics and Automation (ICRA), 2026. (Accepted)
  • I Made Aswin Nahrendra, Byeongho Yu, Minho Oh, Dongkyu Lee, Seunghyun Lee, Hyeonwoo Lee, Hyungtae Lim, and Hyun Myung†, “DreamWaQ++: Obstacle-Aware Quadrupedal Locomotion With Resilient Multi-Modal Reinforcement Learning," IEEE T-RO (Transactions on Robotics), vol.42, pp.819-836, Jan. 2026. [Link]
  • Seunghyun Lee, I Made Aswin Nahrendra, Dongkyu Lee, Byeongho Yu, Minho Oh, Hyeonwoo Lee, and Hyun Myung†, “DreamFLEX: Learning Fault-Aware Quadrupedal Locomotion Controller for Anomaly Situation in Rough Terrains," in Proc. IEEE Int'l Conf. on Robotics and Automation (ICRA), pp. 16001-16007, Atlanta, USA, May 2025. [Link]
  • I Made Aswin Nahrendra, Christian Tirtawardhana, Byeongho Yu, EungChang Lee, and Hyun Myung†, “Retro-RL: Reinforcing Nominal Controller with Deep Reinforcement Learning for Tilting-Rotor Drones," IEEE RA-L (Robotics and Automation Letters), vol.7, no.4, pp.9004-9011, Oct. 2022. [Link]

Photos and Videos