Anti-spoofing Performance Enhancement by Facial Micro-expression Detection using Kinect Sensor
Abhijit Kumar Pal, Debmani Saha · 2021 Sixth International Conference on Image Information Processing (ICIIP) · 2021
Face detection is a point of interest in many systems. Due to the reason of being less-intrusive characteristic. But various systems are not that much capable of preventing high-level facial spoofing attacks. The attacks are generally done by 3D printed masks or eye-cut photos etc. Detecting micro-expressions in this case can make those systems invulnerable to the attacks because micro-expressions are the only expressions that cannot be controlled. In this article, an approach to detect the micro-expressions of the face has been shown. With the successful detection of micro-expression, the liveness of the face can be detected using a histogram of gradient (HOG) descriptor. This descriptor is used to detect the change in pixel intensities. The descriptor has been applied on specific ROI of the face image i.e., two eyes, nose, and lips of the detected face. The dataset utilized in this project is self-created. The device used to capture the snaps of the dataset is Kinect Xbox One. This approach is effective in preventing such face spoofing attacks.