원문정보
초록
영어
In this paper usefulness of time series based Auto Regressive (AR) modelling technique has been explored for identification of a person. For this purpose, time series is obtained from the contour coordinates of the ear. AR model is fitted to this time series. AR coefficients thus obtained serve as a feature vector. Recognition Rate (RR) has been found by a classifier that is based on Euclidian distance between feature vector of test samples with training samples within itself (intraclass) and with respect to others (interclass). Model has been found invariant to posture, rotation and illumination. RR up to 99% has been obtained. Results have been compared with existing techniques. The results demonstrate the effectiveness of technique for human identification.
목차
1. Introduction
1.1. Related Work
1.2. Our Work
2. Edge Detection
2.1. Edge-Detection Technique
3. Time Series Modelling
3.1. Autoregressive Model
4. Implementation
4.1. Test Data
5. Classification
5.1. Co-relation Test
5.2. Recognition Rate
5.3. Performance
6. Comparison with Different Techniques of Ear Biometrics
7. Discussion
8. Conclusion
References