Serious brain injuries, spinal injuries, and motor neuron diseases often lead to severe paralysis. Individuals with such disabilities can benefit from interaction techniques that enable them to interact with the devices and thereby the world around them. While a number of systems have proposed tongue-based gesture detection systems, most of these systems require intrusive instrumentation of the user's body (e.g., tongue piercing, dental retainers, multiple electrodes on chin). In this paper, we propose a wireless, non-intrusive and non-contact facial gesture detection system using X-band Doppler. The system can accurately differentiate between 8 different facial gestures through non-contact sensing, with an average accuracy of 94.3%.