Font Size: a A A

A Fake Iris Detection Algorithm Research Based On Sparse Coding

Posted on:2016-03-05Degree:MasterType:Thesis
Country:ChinaCandidate:J HanFull Text:PDF
GTID:2308330479994726Subject:Computer technology
Abstract/Summary:PDF Full Text Request
With the rapid development of today’s technology, the use of human iris features for individual identification in iris recognition technology has been obtained considerable development. However, in the real complex environment, iris recognition system could been attacked by a lot of fake iris, such as:unrelated category, synthetic iris, print iris and so on. To solve this problem, this paper analyzes those algorithms that related to this studies, and put forward our corresponding algorithms to solve this problem. The main contributions of this paper are divided into the following aspects:1.Established four fake iris databases:synthetic iris database PS1,synthetic iris database PS2,gray-print iris database, color-print iris database. Making two synthetic iris databases PS1 and PS2 through the use of photoshop mask technology, creating two print iris databases by using print+scan mode. These works make great convenience to scholars for later research.2.Do analysis and research on the fake Iris detection algorithm based on image quality assessment algorithm, and pointed out the problems on this algorithm. Pointed out that such fake iris detection algorithms greatly rely on background information on the training set. It cannot be used to detect fake iris under a real complex environment.3.Put forward a fake iris detection algorithm framework based on sparse representation, this algorithm framework mainly includes two parts:a. The SCCR (Sparse Coefficient Concentration Ratio) fake iris detection algorithm is proposed. The largest SCCR value is calculated in the whole training set. Then by setting the threshold value to effectively detect unrelated category, synthesis iris, gray-print iris attacks.b. The color-print iris detection algorithm based on sparse reconstruction is proposed. Through the use of sparse coefficient to reconstruct real and color-print iris image, and then calculate the euclidean distance between the test image and the two reconstructed image. It is successfully solve the problem of color-print fake iris detection.c. Do fusion on this two algorithm, established a fake iris detection system based on sparse representation, it could effectively detect various forms of fake iris attacks.In this paper, through a lot of experiments and analysis show that the fake iris detection algorithm based on sparse representation is able to solve various forms of fake iris attacks.
Keywords/Search Tags:Image Quality Assessment, Fake Iris Detection, Sparse Coefficient Concentration Ratio, Sparse Reconstruction
PDF Full Text Request
Related items