The Research On Iris Location Algorithm Based On Region Of Interest

Iris recognition has become one of the most popular biometric technologys, because of its unique advantages. But the research of iris recogniton has a long way to go, and how to ensure the accuracy of iris recognition is urgently needed to solve.Iris location is a part of image preprocessing and it is the key step in iris recognition. The premise of ensuring the accuracy of iris recognition is ensuring the accuracy of iris location. So, the result of iris location is the key of determining the performance of iris recognition system is good or bad.The thesis introduces the structure of iris recognition sysytem, analyzes and compares the common iris location algorithms in detail, on the basis of these, it presents an iris location algorithm based on region of interest. It proposes to automatically extract the region only containing the complete iris information from an original iris image and remove other useless information, such as shadow, eyebrows and hair, by using Histogram of Oriented Gradients (HOG) to get the statistical information of direction and gradient of an iris image and taking the information into the Support Vector Machines (SVM) for training to get the SVM decision function. In iris inner boundary location, we will divide it into two parts:rough location and accurate location. Firstly, we use the minimum average gray method to locate the innner boundary roughly. Then on the basis of this, we map the annular region on the rectangular region for the iris inner boundary accurate detection. At last, we use the improved J.Daugman algorithm to locate the iris outer boundary.The improved algorithm is used for IrisDatabase CASIA v1.0and CASIA v2.0to analyze the experiment. The result shows that the proposed algorithm can efficiently improve the accuracy of iris location comparing with the common Daugman, Wilds and Least Squares algorithms.