Please use this identifier to cite or link to this item:
|Title:||Robust and Accurate Classical Hough Transform Approach for iris Recognition|
|Keywords:||Hough Transform;Classical Hough Transform|
|Abstract:||As an important part of scientific community, object recognition has become admirably suitable facet of digital image processing. Circular objects are recurrently seen in many images. A large number of methods for circle detection have been studied for various industrial applications. The application province considered in this chapter is ‘Iris recognition’. On the basis of earlier research and results, time comparison and the efficiency of various methods to detect iris in digital images are compared. Iris recognition is one such area which is loaded enough to research into and desires attention and upgrading. It is used in user verification security systems as well as in medical fields. Iris recognition is based on the fact that the significant information for recognition is found within iris outline which is modeled as circle. Rest of the data from the eye region is redundant. Hough transform (HT) is an effective method for confirming the co-ordinates of the center of the circle and its radius. Circle Hough transform (CHT) provides a robust technique for iris detection, but the large amount of storage and computing complexity are the major drawbacks of it. Many modified versions have been applied by researchers in order to reduce computational complexity and memory requirements. Present work proposes a modified CHT which considerably improves the speed of the process without compromising the accuracy of the technique. An exhaustive analysis is conducted for a large number of iris images. In order to achieve the objective, present work is proposing variations in CHT. The work is concerned with the time, computation and memory requirements of CHT. In a more extended version of chapter, information of valid region is exploited to develop appropriate method for speedy and correct extraction of circle. To reduce computation time, HT space is divided into equal sized four quadrants which lie in most probable area of the image to find circle center and radius. These quadrants contribute in finding the significant region of the image which needed to be processed. Hence, determining the valid region is the important step towards fast detection. This also permits limiting the memory needs of the algorithm. After that CHT is applied on all four extracted valid regions which give us fast recognition of the iris.|
|Description:||Master of Technology (Computer Science and Applications)|
|Appears in Collections:||Masters Theses@CSED|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.