| Hi-index | 0.00 |
There are many u-health applications and smart home technologies for a secure, easily applied, low-cost method for identifying individuals for security and medical reasons. However, some biometric systems may not applicable for the disability population where people have lost their hands, eyes, or voice. Personalization is also an essential issue as well because it can adapt users' behaviors, profiles, and patterns. The benefit of a personalized embedded system not only potentially reduces medical device complexity, but also adapts end-user conventionality daily. Once u-health devices are applied on multiple users or on people with disability, an automatic personalization system becomes more essential. The electrocardiogram (ECG) is not only a very useful diagnostic tool for clinical purposes, but is also a new biometric tool for human identification. Best of all, it can be applied on the disability population. ECG biometrics can easily be combined with other biometrics to provide an extra liveness test with little additional cost. A total of 168 normal, healthy individuals were investigated for identification as a predetermined group. Fifty persons were randomly selected from this ECG biometric database as the development dataset. Then, the identification algorithm developed from this group was tested on the entire database. In this research, two algorithms were evaluated for ECG identification during system development. The first is the template matching with LDA neural network model and the second is the fast real-time ECG identification method. The first algorithm provided the identification rate with up to 100% accuracy on the development dataset with an identification rate of 95.3% (160 out of 168 persons) when test database with 168 subjects was applied. The second algorithm reduced processing time, and it is more suitable for portable medical devices. However, the accuracy rate decreased to 89.28%. Hence, Lead-I ECG is a valuable biometric for identifying the disability population. They can be recommended for future biometric systems used for human identification, personalization, and liveness testing.