Research
Research Area
Our group is working on statistical signal processing such as estimation/detection, machine learning, inverse problems and optimization, etc. We apply those methods for machine learning based machine vision system, smart radar signal processing, etc.
Machine learning based machine vision
머신 비전 시스템
Our group is intensively working on developing a machine vision system based on machine learning. Machine vision has been effectively applied for inspecting products during the manufacturing process. Currently, manual inspection is the most common method in factories. We aim to replace the manual methods with fully automated artificial intelligence based on a machine vision system for future smart factories which are the key elements of the 4th industrial revolution. Specifically, we have studied automatic inspection systems for wafer, photomask, and PCB using deep learning architectures such as the Siamese network and segmentation network. Also, we are interested to develop inspection systems using anomaly detection methods, which use only normal data for training networks, because there are little defect data in the real world.?
Smart radar signal processing
- MONITORING PRESENCE OF PEOPLE
- TRACKING MOVEMENTS OF TARGETS
- SIGNAL PROCESSING AND MACHINE LEARNING
Smart radar can be used for monitoring presence of people, bio signal, tracking movements of targets. It can be combined with visual information for better understanding of a scene. We are working on developing statistical signal processing methods for smart radar and trying to apply machine learning methods for understating signals from smart radars.