Our lab focuses on developing intelligent perception, efficient AI systems, and physically grounded AI that interact with the real world.
Computer Vision
We study fundamental and applied problems in visual perception.
- Object detection, tracking, and segmentation
- Feature extraction and matching
- Representation learning for visual understanding
- Vision-based motion and scene analysis
Related R&D Projects
- “The development of 50m range ToF CMOS sensor, optical system and signal processing for automotive” (KEIT, ‘17.04 ~ ‘21.03)
- Demo: Depth-image based object detection (not RGB-image based)
Demo Video — Open in YouTube
- “The development of 4D reconstruction and dynamic deformable action model based hyper realistic service technology” (GigaKorea, ‘17.04 ~ ‘20.12)
- Demo: Progressive 3D human reconstruction based on single RGB camera
Demo Video — Open in YouTube
On-Device AI
We develop efficient AI models that can run on resource-constrained devices.
- Lightweight and efficient neural network design
- Model compression, quantization, and pruning
- Real-time inference on mobile and edge devices
- Deployment-aware training and optimization
Related R&D Projects
- “Development of intelligent media aspect ratio conversion technology that maintains properties” (IITP, ‘21.04 ~ ‘24.12)
- Demo: Real-time image super-resolution on mobile devices
Demo Video — Open in YouTube
Demo Video — Open in YouTube
Digital Human & Sports AI
We aim to model, analyze, and simulate human motion and behavior.
- Human pose estimation and motion analysis
- Digital human modeling and simulation
- Personalized exercise and sports training AI
- AI-based performance analysis and feedback
Related R&D Projects
- “Development of metaverse-based sports play tweening and realistic technology” (KOCCA, ‘22.04 ~ ‘24.12)
- Demo: Real-time 3d skeleton extraction for golf swing
Demo Video — Open in YouTube
Physical AI
We study AI systems that perceive, reason, and act in the physical world by tightly integrating perception, decision-making, and control.
- Vision-based robotic perception and state estimation
- Learning-based control and decision-making
- Embodied AI and sensorimotor learning
- Human–robot interaction and physical intelligence
- Integration of Computer Vision and On-Device AI for real-world robotics
Related R&D Projects
- “Preparing”