| Title | Insider Threat Detection with Face Recognition and KNN User Classification |
| Publication Type | Conference Paper |
| Year of Publication | 2017 |
| Authors | Sarma, M. S., Srinivas, Y., Abhiram, M., Ullala, L., Prasanthi, M. S., Rao, J. R. |
| Conference Name | 2017 IEEE International Conference on Cloud Computing in Emerging Markets (CCEM) |
| Keywords | authentication, authentication mechanism, authorisation, cloud computing, cloud insider, Cloud Penetration, cloud storage, cloud user community, Degree of Trust, Face, face recognition, Facial features, factor authentication, Human Behavior, image authenticity, Information security, insider threat, Insider Threat Detection, insider threats, kNN classification algorithm, KNN Classification and QoS, KNN user Classification, machine learning, message authentication, Metrics, Monitoring, nearest neighbour methods, pattern classification, policy-based governance, private cloud deployments, pubcrawl, quality of service, resilience, Resiliency, security concerns, Security Method, security of data, sensitive information, threat detection model, Threat detection module, threat detection QoS |
| Abstract | Information Security in cloud storage is a key trepidation with regards to Degree of Trust and Cloud Penetration. Cloud user community needs to ascertain performance and security via QoS. Numerous models have been proposed [2] [3] [6][7] to deal with security concerns. Detection and prevention of insider threats are concerns that also need to be tackled. Since the attacker is aware of sensitive information, threats due to cloud insider is a grave concern. In this paper, we have proposed an authentication mechanism, which performs authentication based on verifying facial features of the cloud user, in addition to username and password, thereby acting as two factor authentication. New QoS has been proposed which is capable of monitoring and detection of insider threats using Machine Learning Techniques. KNN Classification Algorithm has been used to classify users into legitimate, possibly legitimate, possibly not legitimate and not legitimate groups to verify image authenticity to conclude, whether there is any possible insider threat. A threat detection model has also been proposed for insider threats, which utilizes Facial recognition and Monitoring models. Security Method put forth in [6] [7] is honed to include threat detection QoS to earn higher degree of trust from cloud user community. As a recommendation, Threat detection module should be harnessed in private cloud deployments like Defense and Pharma applications. Experimentation has been conducted using open source Machine Learning libraries and results have been attached in this paper. |
| DOI | 10.1109/CCEM.2017.16 |
| Citation Key | sarma_insider_2017 |
Insider Threat Detection with Face Recognition and KNN User Classification