Daegu Gyeongbuk Institute of Science and Technology

Hui Sup Cho

Young Jin Park

A motion recognition method may include receiving a plurality of frame signals based on radar pulses reflected from a target at different times, generating a plurality of micro-range enhanced frame signals obtained by reinforcing a component of a second region having more movement than movement of a first region among the plurality of frame signals, generating a micro-range enhanced frame set formed by stacking the plurality of micro-range enhanced frame signals at preset time intervals, generating a micro-range enhanced radar image which is an image obtained by viewing the micro-range enhanced frame set in a direction perpendicular to a time axis of the micro-range enhanced frame set, and determining motion of the target by inputting the micro-range enhanced radar image to a machine learning-based learning model.

Apparatus and method for motion recognition based on pulse radar

An apparatus and method for determining bio-information of a target using an impulse radar are provided. The method may include generating a frame set by accumulating frames received at preset time intervals, determining a first magnitude spectrum of the frame set corresponding to a first frequency axis by performing a frequency conversion of frames included in the frame set in a sampler index axis direction, determining a second magnitude spectrum of the frame set corresponding to the first frequency axis and a second frequency axis by performing a frequency conversion of the first magnitude spectrum in a time axis direction, determining a third magnitude spectrum of the frame set by adding up values of the second magnitude spectrum for each second frequency, and determining a frequency indicating a peak in the third magnitude spectrum as a heartbeat frequency of the target.

Growing community of inventors

Daegu, South Korea

Hui Sup Cho