We invite you to the 6th seminar of 'SNU A.I. MED Talks 23'.
This seminar will be presented by Professor Kim, Youngwook (Department of Electronic Engineering) about
"Applications of Electromagnetics for Analyzing Human Motions Using Deep Learning".
- Date: October 19th, 2023. (Thur.) 16:00-17:00
- Venue: (via Zoom) https://snu-ac-kr.zoom.us/j/
- Topic : Applications of Electromagnetics for Analyzing Human Motions Using Deep Learnin
Abstract :
This presentation discusses the applications of electromagnetics for monitoring human subjects using Deep learning algorithms. Radar has been employed for defense purposes for several decades, but recently, it offers a new opportunity for application to human subjects. This presentation covers three main topics. Firstly, it addresses human detection and monitoring problems using radio frequency sensors. Human monitoring has been an important research area for a wide variety of applications, including security, surveillance, border patrol, and search-and-rescue missions. To process the sensor data, machine learning algorithms, such as the support vector machine and deep convolutional neural network, are suggested to detect humans and analyze their activities. In particular, the feasibility of using generative adversarial network for the synthesis of radar image is addressed. Secondly, applications to the human body for health care (on-body biometrics) are discussed. Sensing vial signals and extracting characteristics of the human body are unique and significant areas of applied electromagnetics. The movement of body parts, such as limbs, heart, eyelids, and fingers, can be measured by Doppler sensors. By investigating micro-Dopplers from human parts, useful information can be induced, including estimation of total energy expenditure, heart rate detection, eye-blinking detection, hand gesture recognition, and even voice. This technology offers the potential to remotely monitor human health and could be applied to health care systems, such as telemedicine and e-health. Finally, this presentation briefly discusses applications of on-body channels and antenna sensors to address human monitoring problems. Through on-body channels and antenna impedance variations, which can be readily used by hand-held devices, human activities from large motions to fine movements are investigated.