Development of Image Classification Model on Face Spoofing Attack Using Ensemble Learning Method

Face spoofing is an attack on a biometric system using a fake identity from a user with access. This attack can be carried out using a printed photo of the user's face or a replay attack using another device that displays the user's face. The prevention method for this attack can be done by detecting incoming data via the biometric system. Therefore a system that can better predict face spoofing is needed. This research will use two machine learning methods: Support Machine Vector (SVM) and Bagging with SVM. The dataset was collected from nine different people and consisted of five different categories there are actual image, print attack, replay attack, mask attack and mask attack with a hole in the eye. After going to the preprocessing phase and model training using the dataset the following accuracy SVM method without the Bagging method was obtained for 91%. The new accuracy was obtained after adding that method as a base estimator into the Bagging method for SVM-Bagging 93%.