Multimodal Biometric Authentication System using ECG Signals and Fingerprints

In some of the aspects especially on security and protection it becomes important that the threats of digital crimes are addressed. Biometrics in general, which at the initial stage are pattern recognition systems, scan the unique characteristics using optical or electronic scanners. However, Non-university and noises in sensed data help in achieving higher GAR and FAR (False Acceptance Rates). Privacy and security are major issues in collecting and storing sensitive biometric data. Strict safeguards are required to prevent unauthorized access or breaches of Electrocardiography (ECG) data due to its sensitive nature and the fact that it is unique to each individual. Hence as solutions to these problems, a multimodal biometric system combining both the ECG signals and the fingerprint signals are suggested. Alterations to electrocardiogram signals over time can be caused by things like getting older or your health getting worse. An enormous obstacle is the fact that current systems can't consistently handle these long-term fluctuations. The proposed new deep architecture for multimodal learning integrates the feature vectors in a stacking as well as channel-wise method. One ought to notice that, during the course of the authentication process, Random Forest Classifier builds a number of decision trees. Knowing that the multimodal biometric system has higher accuracy and resisting power than the unimodal one, but at the same time has a number of problems embedded, when switching between them it is crucial.