ACADSTAFF UGM
| Title | : | Multimodal Biometric Model Development with Hand Geometry and Voice Features |
| Author | : |
PATRICIA EVERICHO M. (1) Prof. Drs. Agus Harjoko, M.Sc., Ph.D. (2) |
| Date | : | 0 2022 |
| Keyword | : | Biometrics,Multimodal,handgeometry,voice Biometrics,Multimodal,handgeometry,voice |
| Abstract | : | The concept of security evolves over time. Strong authentication is increasingly being improved. The methods used to ensure the integrity of user authentication of a device are also growing. In order to increase the level of security, the biometrics model is considered suitable to replace the conventional security model. Besides being able to support the high mobility of modern society, biometrics are also able to ensure that the only person who can open the lock is the owner of the device. The purpose of this study is to develop a multimodal biometric model that combines hand geometry and voice features. Compatibility of the biometrics is obtained by calculating the Euclidean distance on the hand images and voice signal. The comparison between testing data and reference data is done after extracting all the features. The features comparison data will be used in the testing process until matching results are obtained. Matching results of each biometrics feature are then normalized according to the min-max rule and combined with the weighted-sum rule score-level fusion. The fusion scores are then sorted out and the lowest value is searched to bring the decision. The test results show that the multimodal biometric model always has the lowest EER value when compared to the unimodal model, which is 8.89% for a total of 30 tests; 7.78% for a total of 25 tests, 6.11% for a total of 20 and 15 tests; 6.67% for a total of 10 tests; and 2.22% for a total of 5 tests. The proposed multimodal biometric model has been successfully developed and proven to be more reliable than its unimodal models. |
| Group of Knowledge | : | Ilmu Komputer |
| Original Language | : | English |
| Level | : | Internasional |
| Status | : |
Published
|