Active Face Spoof Detection Using Image Distortion Analysis

With the rising use of facial recognition systems in a range real-world scenarios and applications, attackers are also increasing their efforts, with number spoofing techniques emerging. As result, developing reliable spoof detection mechanism is critical. Active-based have been shown to be good at finding spoofs, but they problems, such as being intrusive, expensive, hard compute, not able used many situations, usually needing extra hardware. This research presented an active-based robust technique capable detecting wide media or 2D attacks while less low complexity, more generalizable than other techniques. It doesn't require any additional hardware, so it can easily integrated into current systems. The distortion variations video frames user's face collected varying distances from camera analyzed detect spoofing. Both legitimate attack datasets were created using photo data. proposed approach achieved accuracy 98.18% both machine learning classifiers deep model, equal error rate half total 0.023 0.021, respectively.