| Summary: | 碩士 === 國立中興大學 === 資訊科學與工程學系所 === 97 === In this report, we proposed energy factors method analyzes the iris information to distribute and to apply on individual identity confirmation system. When image pre-processing, we obtain iris region information and normalized to a fixed measurable region, after standardization, will carry out second-order wavelet transform and remove non-prominent features. This process will then allow us to carry on re-distribution work and transform these regions to energy factors state. Calculation will be performed on each region’s energy charact-eristic, and this region’s energy content will then be characterized as iris’s finger prints.
Utilize these energy characteristics to produce a group of revolving gene. Revolving gene can then be used to compare and make the necessary adjustments. The differences in energy characteristics between two irises can also be used to gauge the similarity of the two. Dual threshold standard can then be applied to examine whether irises fall into the same category.
In this report, we carried out system testing on 108 groups and for the total of 756 iris images tested, the proposed method have the success rate of 98.02% in iris identification and equal error rate of 11.69%. In the case of FAR=0 also resulted in low FRR value of 22.97. Hence, this validated our proposed method of using energy factors expressed in iris texture region distributed information. As the result, this process proves to be highly efficient in recognition rate and also under the high security requirements situation.
|
|---|
