Intrusion detection based on biometric spoofing and network anomaly detection with machine learning algorithms

In this study, we proposed four different approaches for intrusion detection in systems. Based on the information, every online platform can attack in different ways. In this project, we used deep learning algorithms for biometric fraud detection. We also proposed a new approach based on anomaly detection in networks for different attacks such as DDOS. In this case, our project is divided into two main parts: Face spoofing detection: Face spoofing attacks are one of the intrusion detection attacks. We have proposed two new applications for face spoofing detection, which we divided into two different groups. The first one, we proposed IoT-cloud based platform for face spoofing detection with deep multicolor feature learning. The second method suggested for face spoofing detection is based on motion analysis with the help of Robust principal component analysis and deep belief network. Network intrusion and anomaly detection: In this section, we have reviewed and suggested two different methods for network attack detection. The third one, we proposed the Combination of Feature Selection and Fast Learning Network based on the Multipurpose Particle Swarm Algorithm. The forth one, We proposed Anomaly and Signature Based IDS for Network Security Using Hybrid Inference Systems.