Detection of reflection in iris images using back propagation

作者：Thenmozhi S; Balasubramanie P; Venkatesh J^*; Aarthy C

来源：Life Science Journal-Acta Zhengzhou University Overseas Edition, 2012, 9(3): 2446-2450.

• Image Processing
• detection
• localization
• feed forward
• back propagation

In our work with the help of back propagation and RST algorithm reflection in iris images are located and removed. From the database the biometric templates from iris images are extracted which are later used for comparison, this phase is called Iris encoding. A pattern recognition system with 2 folds is present in order to recognize the image pattern; extract the set of features from iris using RST [Radial symmetry Transform] algorithm; an inductive classifier to achieve reflection, detection and location by a pixel by pixel analysis. The use of the FFNN method instead of the RSBIN or BIN method can achieve a more robust behaviour of the system with respect to the threshold value. The whole system has been implemented in Mat lab by exploiting the available Neural Network Toolbox and PROTOCOL. The project description consists of five modules. An iris image is extracted, reflection is detected and then the feature extraction, selection and fusion is carried over with classification system and error is estimated. The RST used focus on neural network as classifiers and this RST is best suited for detection and localization of reflection. [Thenmozhi. S, Balasubramanie. P, Venkatesh. J, Aarthy. C. Detection of reflection in iris images using back propagation. Life Sci J 2012;9(3):2446-2450] (ISSN:1097-8135). http://www.lifesciencesite.com. 353

• 出版日期2012