Research Interests

Computer Vision

At the Smart Vision & Media Lab., we conduct cutting-edge research in computer vision with applications in content generation, autonomous driving, 3D human modeling and analysis, gesture recognition, and visual inspection. Specifically, our work includes video generation using generative models, multi-camera autonomous driving based on object detection, 3D human modeling and analysis using point cloud data, action recognition including hand gestures, and vision inspection technologies for manufacturing.

Training Methods for Deep Learning Models
In industry, collecting labeled training data is often a major challenge, which can limit the effective training of deep learning models. To address this issue, our lab focuses on semi-supervised learning and domain adaptation techniques that enable high performance even with limited data.

Optimization of Deep Learning Models
We develop computer vision and image processing algorithms designed for real-time systems such as display technologies. To achieve efficient deployment, we study model compression and optimization techniques including pruning and quantization, with the goal of ensuring real-time processing performance.