Biometric Authentication for Vehicle Access and Driver Stress Detection
Biometrics authentication has been widely applied in many domains and has gained even more popularity with the advances of smart devices. The traditional methods such as passwords, key and lock systems can no longer keep up with the demand for more secure and convenient mode of user authentication. Therefore, this research work aims to explore the possibility of using multimodal biometrics using electrocardiogram (ECG) and fingerprint to replace the traditional key system in smart vehicles. This end-to-end deep neural network makes use of the siamese VGG16 architecture to extract features and performing learning from ECG and fingerprint separately. The learnt features from both ECG and fingerprint are then combined at the second last fully connected layers to perform classification. The highest accuracy obtained from user verification task is 97.84% which is comparable to the state-of-the-art results. This work further explored the use of ECG as a method for identifying driver stress which can be used to mitigate the possibility of road rage accidents and obtained an accuracy of 76.34%.
L. Zhiyuan, "Biometric Authentication for Vehicle Access and Driver Stress Detection", M.S. Thesis, Machine Learning, MBZUAI, Abu Dhabi, UAE, 2022.