Detection of Eye Blinking Using Doppler Sensor With Principal Component Analysis

We propose the detection of human eye blinking using a Doppler sensor. Eye blinking is one of the most common and convenient human activities that can be used as an input modality for computer devices, a simple communication methodology, and fatigue diagnostics. The reflected wave from the blinking eye has a unique Doppler signature. To investigate this signature, several measurements were performed with/without the noise caused by human movement when the sensor was placed near the eyes. We analyzed the Doppler signal in the joint time-frequency domain. It was found that the Doppler frequency produced by eye blinking is approximately 115 Hz. Furthermore, unconscious and conscious eye blinking exhibited different Doppler characteristics. In order to classify these characteristics, we employed a principal component analysis to extract the features. The truncated eigenvectors were multiplied with an image of eye blinking, and the resulting coefficients were used for classification. The proposed method successfully differentiated conscious eye blinking in the presence of noise from human motion in diverse scenarios.