Disclosed are a bio-signal-based eye tracking system using a dual machine learning structure and an eye tracking method using the same. The bio-signal-based eye tracking system using a dual machine learning structure comprises: a bio-signal measuring portion (110) for detecting a bio-signal produced by a movement of eyes moved in accordance with the movements of user eyes; a signal separating portion (120) for separating the bio-signal into a safety signal and an electromyographic signal in accordance with a frequency range; a safety signal processing portion (130) for simplifying a wave form of the safety signal into a plurality of consecutive segments by applying to a PLA method, listing the segments as a feature set having a multi-dimensional vector shape, and outputting a first process signal including distance information in accordance with the movement of user eyes in accordance with the movement of eyes by applying a dual machine learning to the feature set; an electromyographic signal processing portion (140) for simplifying a wave form of the electromyographic signal into a plurality of consecutive segments by applying to the PLA method, comparing an amplitude value of the electromyographic signal of a temporal muscle of a point that a user bites with a threshold value preset inside between the segments, and outputting a second process signal for carrying out a mouse click if the amplitude value is greater than the threshold value; and a mouse driving portion (150) for moving a mouse cursor by responding to the distance information of the first process signal, and operating a click button of a mouse by responding to the second process signal.
展开▼