The appearance of human eyes consists of sclera, iris and pupil. The sclera is the white part around the eyeball, accounting for about 30% of the total area; The center of the eye is the pupil, accounting for about 5%; Iris is located between the sclera and the pupil, and contains the richest texture information, accounting for 65%. Externally, it is composed of many glandular pits, folds and pigment spots (see the right), which is one of the most unique structures in the human body. The formation of iris is determined by genetic genes, and human gene expression determines the morphology, physiology, color and overall appearance of iris. When a person develops to about eight months, the iris is basically large enough to enter a relatively stable period. The appearance of the iris can remain unchanged for decades, unless there are rare abnormalities and great physical or mental trauma. On the other hand, the iris is externally visible, but at the same time it belongs to the internal tissue and is located behind the cornea. Changing the appearance of the iris requires very delicate surgery and there is a risk of visual damage. Iris has a high degree of uniqueness, stability and invariance, which is the material basis of iris recognition.
Among all biometric technologies including fingerprints, iris recognition is the most convenient and accurate one. Iris recognition technology is widely regarded as the most promising biometric authentication technology in 2 1 century, and its future application in security, national defense, e-commerce and other fields will inevitably focus on iris recognition technology. This trend is gradually emerging in various applications around the world, and the market application prospect is very broad.
Foreign research institutions of iris recognition mainly include Iridian and Iriteck in the United States and Jiris in South Korea. Iridian Company has mastered the core algorithm of iris recognition, and is currently the largest professional iris recognition technology and product provider in the world. It cooperates with LG, Panasonic, OKI, NEC and other enterprises (such as IRISPASS? 、BM-ET300、IG-H 100？ Products such as) provide the core algorithm of iris recognition in an authorized way, and support partners to produce iris recognition systems. Iridian's core technology also includes image processing protocol and data standard PrivateID? , identify the server KnoWho? , who knows? Develop tools and iris recognition cameras.
Before 2000, China did not have its own core intellectual property rights in iris recognition. Based on years of research, the Institute of Automation of China Academy of Sciences developed the core algorithm of iris recognition in early 2000, becoming one of the few units in the world to master the core algorithm of iris recognition.
As the first research base in China to study the mechanism of iris recognition, the state key laboratory of pattern recognition, Institute of Automation, Chinese Academy of Sciences, has developed the iris recognition in vivo detection technology with independent intellectual property rights, which not only fills the gap in the international field of iris recognition technology in China, but also can compete with the international mainstream algorithms. Compared with the core algorithms of other international companies, the core algorithms of Institute of Automation of Chinese Academy of Sciences are faster, occupy less memory space and have better overall performance. In 2005, the research results of iris recognition in our laboratory won the second prize of national scientific and technological invention. In September 2006, the State Key Laboratory of Pattern Recognition, as the authoritative organization of iris recognition technology in China, participated in the evaluation of biometric technology organized by international biometric organizations (2006 Biometric Alliance Conference and 2006 Biometric Technology Experiment), and its speed and accuracy of iris recognition algorithm were recognized by international peers. In addition, the iris image database of the State Key Laboratory of Pattern Recognition has become the largest iris database in the world. Up to now, more than 2,000 research institutions in 70 countries and regions have applied to use it, including more than 1500 foreign institutions.
Beijing Zhongke Hongba Technology Co., Ltd. is a company specializing in the research and development of iris recognition system and related products based on the scientific research results of iris recognition in the State Key Laboratory of Pattern Recognition. It is the only enterprise in China with completely independent intellectual property rights of iris recognition technology. * * * Company obtained iris patents 15, including 8 invention patents, 4 utility model patents and 3 appearance patents, and obtained software copyright. Intellectual property covers all aspects such as iris recognition software and hardware systems. At present, Zhongke Hongba's iris recognition products have been successfully applied in some domestic coal mines, banks, social security and high-end classification places.
The automatic iris recognition system includes two modules: hardware and software: iris image acquisition equipment and iris recognition algorithm. Corresponding to two basic problems: image acquisition and pattern matching.
1993, JOHN DAUGMAN realized a high-performance automatic iris recognition prototype system. Today, most automatic iris recognition systems use DAUGMAN core algorithm.
Iris is a ring-shaped part between the black pupil and the white sclera of the eye, which is generally radial from the inside out. It is composed of quite complex fibrous tissues and contains many interlaced details, such as spots, filaments, crowns, stripes, depressions and so on. These characteristics are determined by the random combination before birth, and once formed, they will remain unchanged for life. The accuracy of iris recognition is the highest among all kinds of biometric recognition.
Information collection:
According to the iris with the diameter of 1 1mm, Dr. Daugman's algorithm uses 3.4 bytes of data to represent the iris information per square millimeter, so that an iris has about 266 quantitative feature points, while the general biometric technology only has 13 to 60 feature points. Iris recognition algorithm with 266 quantized feature points is described in many iris recognition technical data. If the algorithm and human eye characteristics allow,
Dr. Daugman pointed out that his algorithm can obtain 173 independent feature points with binary degrees of freedom. In biometric technology, the number of this feature point is quite large.
Algorithm:
The first step is to locate the iris with a precision camera 3 inches from the eye. When the camera is aimed at the eyes, the algorithm gradually focuses on the left and right sides of the iris to determine the outer edge of the iris. This horizontal method is hindered by eyelids. At the same time, the algorithm aims at the inner edge of iris (that is, pupil) and eliminates the influence of eye fluid and fine tissue.
Monochrome cameras use visible light and infrared light, and infrared light is in the range of 700-900mm (this is the lower limit of IR technology, and the American Ophthalmology Association used the same range in their research on macularcysts. )
Above the iris, the algorithm uses two-dimensional gabor wavelet to subdivide and recombine the iris image. The first subdivision is called phasor, which requires deep mathematical knowledge to understand the principle of two-dimensional Gabor wavelet.