Mitigating Face Biometric on Electronic Medical Record

With the current movement of patient filing system from paper-based system towards Electronic Medical Record (EMR) system, it is necessary to protect patient privacy on EMR efficiently. This research incorporate faces biometric system into EMR in order to provide a twolayer security for patient confidential information. However, the issue of face with distortion such as expressions, aging, background illumination etc. which often affects the performance of face biometric system during authentication stage was addressed. We utilized optimized Principal Component Analysis and Support Vector Machine (PCA-SVM) to extract features and address this issue of patient’s face with distortions. We conducted an experiment with 100 test cases. This result shows that facial biometrics can be used for securing patient electronic medical record in hospitals based on the system accuracy obtained. The result demonstrated 78% system performance accuracy with distorted face, which shows that optimized PCA-SVM exhibit a reliable performance for promising security work in term of False Rejection Rate(FRR) and False Acceptance Rate (FAR).