A Compute-Efficient Algorithm for Robust Eyebrow Detection

Detection of facial landmark features is an initial step in facial expression recognition. Detection of eyebrows can aid the detection of the remaining facial features as eyebrows are relatively more stable across changing facial expressions. Existing eyebrow detection algorithms in literature involve complex computations and are not suitable for direct porting on to embedded platforms. In this paper, a compute-efficient eyebrow detection algorithm has been proposed and tested on three standard databases with an average detection rate of 96%. The method has also been shown to be computationally less complex compared to the state of the art.