Abstract
Conference Title: 2018 IEEE 4th International Conference on Identity, Security, and Behavior Analysis (ISBA) Conference Start Date: 2018, Jan. 11 Conference End Date: 2018, Jan. 12 Conference Location: Singapore, Singapore Biometrie data plays an important role today as an identity authentication tool. However, designing an efficient and secure biometrics authentication scheme in the cloud environment remains a research challenge. To address the security issue ofbiometric data in the cloud, we propose Cloud-ID-Screen scheme. Cloud-ID-Screen utilizes low-cost cloud storage and scalable cloud computation to improve security and privacy, as well as process speed. Cloud-ID-Screen achieves this improvement by splitting fingerprint features into small subsets and distributing these small subsets into multiple clouds simultaneously, where no single cloud stores all subsets of a fingerprint. During the matching operation, Cloud-ID-Screen utilizes a parallel processing tool (e.g. Hadoop) to match each subset of the fingerprint features independent and in parallel to maintain security and improve process speed. Our experiments show that Cloud-ID-Screen achieves comparable accuracy in matching while being statistically significantly faster as compared to the baseline.