| Summary: | 碩士 === 國立臺北科技大學 === 電子工程系研究所 === 103 === Recently, due to the advance of wearable devices and wireless sensoring technologies, we can measure many physiological signal timely. Including blood pressure, blood oxygen levels and sleep quality as well as fingerprints, iris, retina, and ECG etc., used in the identity authentication. And now current common identity authentication such as: 1. Fingerprint authentication have high and low cost advantages, but there have some artificial factors make fingerprint easily replicated; 2. Iris authentication usually using some characteristics around the crystalline structure as a template, but it need bright enough and difficult to read for a black eye; 3. Retinal authentication usually using visible light or infrared to get the features at the back of eyes as a template, but it may harm users&;#39; eyes; ECG authentication using one lead ECG to do the personal identity,but it&;#39;s not enough to capture ECG characteristics unless using the 12-lead ECG to increase the number of features. If we do ECG authentication individually, it will fail to identify owing to the lack of features. ECG signals will changed by age and mood.
In this paper, we using the fingerprint and electrocardiogram as a multiple authentication for a personal identity. There have identification rate of 96% in fingerprint authentication.In ECG authentication, there have 92.6% using fixed threshold method and 95.9% using variable threshold method. After a multiple authentication, it has identification rate of 92%. Compared with other research, we reduce the complexity of algorithms, but still able to achieve a nice identification rate, and it will allow to use in the wearable device.
|
|---|
