CLIAMS:1. A method for detecting liveness in iris of an eye, the method comprising steps of:
capturing a first enrolled iris template images without illuminating the eye followed by a second enrolled iris template images by illuminating the eye with LED lights source to form a data base;
decomposing the first and second enrolled iris template images;
extracting information from activity bands (1 LL) of the first and second enrolled iris template images for capturing epigenetic response of human eye toward the LED light of predefined frequency by comparing the first and the second enrolled iris template image;
sub-level decomposing of the first and the second enrolled iris template images for pattern recognition;
filtering the first and the second enrolled iris template image for reconstructing the first and the second template iris image to obtain seven high bands thereby obtaining energy distribution pattern and detecting conjunctival vascular pattern for developing liveness in enrolled iris template images ;
storing the first and the second enrolled iris template images in the database for further reference;
capturing a first and second verification iris images and passing through steps of decomposing, extracting, sub-level decomposing and filtering thereby detecting conjunctival vascular pattern for developing liveness verification iris images; and
matching a first and the second verification iris image with the first and second enrolled template iris image to identify perfect match and for recognition and determining liveness.

2. The method as claimed in claim 1, wherein the second iris images are captured by illuminating infra-red light with wavelength in the range of 790nm to 850nm.

3. The method as claimed in claim 1, wherein the decomposition is done by using Daubechies wavelet-based approach.

4. The method as claimed in claim 1, wherein the time interval between the first iris image and the second iris image is of 500 mili-seconds.

5. The method as claimed in claim 1, wherein wavelet packets are used for extracting information from activity bands (1 LL ) of the iris images.

6. The method as claimed in claim 1, wherein the iris image is filtered by orthogonal wavelet filters.
,TagSPECI:Field of the invention

The present invention relates to a method for detecting liveness in iris of an eye.

Background of the invention

Biometric information is extremely sensitive and is required to be protected while operating in real situations. It is thus very much desired that biometric systems are secured and matured before they are deployed in real-life situations. Iris patterns verification is one of the methods used for biometric identification.

Further, iris of an eye is considered as an ideal part of the human body for biometric identification for following reasons:

Iris is an internal organ that is well protected against damage and wear by a highly transparent and sensitive membrane (the cornea). This distinguishes it from fingerprints. Fingerprints can be difficult to recognize after years of certain types of manual labor causing change in the pattern. Furthermore, iris is mostly flat, and its geometric configuration is only controlled by two complementary muscles (the sphincter pupillae and dilator pupillae) that control the diameter of the pupil. This makes the iris shape far more predictable than, for instance, that of the face. Iris has a fine texture like fingerprints, which determined randomly during embryonic gestation.

There are iris recognition methods at present, but these methods have some major drawbacks and can be spoofed by the following ways:
1. A high resolution photograph of eye when kept in front of scanner, the method cannot differentiate it from real eye.
2. Artificial eye of person can be made out of plastic or glass, can fool the method.
3. Eye of the person removed from body is taken as a live eye of the person by the method.

Therefore, there is a need to provide an iris recognition a method, which comes all the drawbacks of the iris recognition systems of the prior art, we have made an extension to it by adding “liveness” to this system.

Objects of the invention

Object of the present invention is to provide an iris recognition method, which are able to differentiate between the image and iris of a live eye.

Another object of the present invention is to provide an iris recognition method, which are able to differentiate between iris of a plastic eye and iris of a live eye.

Yet another object of the present invention is to provide an iris recognition method, which is able to detect eye that is separated from body.

Summary of the invention

According to the present invention there is provided a method for detecting liveness in iris of an eye. The method comprising steps of capturing a first enrolled iris template images without illuminating the eye followed by a second enrolled iris template images by illuminating the eye with LED lights source to form a data base. Further, decomposing the first and second enrolled iris template images. Thereafter, extracting information from activity bands (1 LL) of the first and second enrolled iris template images for capturing epigenetic response of human eye toward the LED light of predefined frequency by comparing the first and the second enrolled iris template image. After that, sub-level decomposing of the first and the second enrolled iris template images for pattern recognition. Further, filtering the first and the second enrolled iris template image for reconstructing the first and the second template iris image to obtain seven high bands thereby obtaining energy distribution pattern and detecting conjunctival vascular pattern for developing liveness in enrolled iris template images. Thereafter, storing the first and the second enrolled iris template images in the database for further reference. After that, capturing a first and second verification iris images and passing through steps of decomposing, extracting, sub-level decomposing and filtering thereby detecting conjunctival vascular pattern for developing liveness verification iris images. At last matching a first and the second verification iris image with the first and second enrolled template iris image to identify perfect match and for recognition and determining liveness.

Brief description of the invention

Figure 1 shows a flow chart for a method for detecting liveness in iris of an eye in accordance with the present invention;

Figure 2 shows 2 shows iris images captured in time sequence of 500 mili-seconds;

Figure 3 shows Shannon entropy bases wavelet packet analysis of the iris images, and

Figure 4 shows characterization of the live-iris patters.

Detail description of the invention

For a thorough understanding of the present invention, reference is to be made to the following detailed description, including the appended claims, in connection with the above-described drawings. Although the present invention is described in connection with exemplary embodiments, the present invention is not intended to be limited to the specific forms set forth herein. It is understood that various omissions and substitutions of equivalents are contemplated as circumstances may suggest or render expedient, but these are intended to cover the application or implementation without departing from the spirit or scope of the claims of the present invention. Also, it is to be understood that the phraseology and terminology used herein is for the purpose of description and should not be regarded as limiting.

The term such as “first” and “second” does not denote priority or any sequence, but are used for differentiating two similar elements. Further, the term “a” and “an” herein do not denote a limitation of quantity, but rather denote the presence of at least one of the referenced item.

The present invention provides a method for detecting liveness in an iris of an eye. The method can differentiate between the image and iris of a live eye. Further, the method enables to differentiate between iris of a plastic eye and iris of a live eye. Moreover, the method enables to detect eye that is separated from body.

Referring now to figure 1, a flow chart of a method (100) for detecting liveness in an iris of an eye is illustrated. The method (100) includes steps of:

At step (10) the method (100) starts.

At step (12), a first enrolled iris template image of an eye of a human which needs to be verified is captured using an iris scanner and the second enrolled iris template image is captured illuminating the eye by LED light. In an embodiment, the LED light with wave length in the range of 790 to 850 nm is used. For the purpose of explanation, the LED light with wave length 795 nm is used. Specifically, epigenetic reflexes of the eyes are evaluated. If it is considered that the first enrolled iris template image is taken at zeroth second, than the second enrolled iris template image is captured after illumining the eye with LED light after time interval of 500 mili-seconds. Figure 2 shows images captured in time sequence of 500 mili-seconds.

At step (14), the first and the second enrolled iris template images are decomposed by using Daubechies wavelet based algorithm. Daubechies wavelets are a family of orthogonal wavelets defining a discrete wavelet transform and characterized by a maximal number of vanishing moments for some given support. With each wavelet type of this class, there is a scaling function (called the father wavelet) which generates an orthogonal multi-resolution analysis.

At step (16), wavelet packets are used to extract information from activity band, which are decided using entropy formula from the first and the second enrolled iris template images. Activity area of more than 70% entropy change is further used for nodal analysis. Typically, the algorithm captures the epigenetic response of human iris when stimulated with a known and carefully selected wavelength of infra red LED light source. In an embodiment, the LED light source is an infra red LED light source. The retained coefficient relates directly to the live-iris pattern.

At step (18), the first and the second enrolled iris template images are subjected to sub-level decomposition for recognition of patterns.

Assuming existence of nested sequences of subspaces {Vj}j8 = - 8 the selected set of Dacubechies scaling function is {F(x –k)}k?z is an orthogonal bases i.e.

(1)

Table 1
Data Set Distribution

Live L1ID CrossMatch LG Time
Inter-class 1456 1456 1456 5 months
Intra-class 169728 169728 169728 5 months

For the vector space Vj spanned by the discrete scaling function:

and the set constitutes the bases.

Now and Vj+1 has better refinement than Vj. This “difference” is subset of Vj+1 spanned by the discrete wavelet of the subspace Wj, gj (x) ? Wj,

The basic of this vector space Wj are always orthogonal to scaling functions on ,

At step (20), the first and the second enrolled iris template images are filtered for reconstruction these images to obtain seven high bands thereby obtaining energy distribution pattern and detecting conjunctival vascular pattern for developing liveness in enrolled iris template images.

The wavelet filters and from an orthogonal filter set if they satisfy following conditions,

Where, stands for low pass and high pass decomposition filter, while are reconstruction filter. Low pass filter come form ? functions, while high pass filters come from ? functions. If the base images is denoted by , then applying both low as well as high filters in horizontal and vertical directions would result in four sub-bands, namely LL, HL, LH and HH, they can be written as,

Here, is the low frequency component, are high frequency components in horizontal, vertical and diagonal direction, respectively, after decomposing the image. Since, the filter used is orthogonal filter, as they follow properties in equation (1), the original image can be reconstructed from these sub-bands using following reconstruction formulae below.

Above referred formula, after extending it to decomposition for wavelet packet analysis, are used to visualize the live-iris pattern, after doing the thresholding. By the end of the process we will have 1 LL (Low pass band for scale 1) band and seven high pass bands as shown in figure 3. Sub-band alignment is performed as the complete algorithm itself is sub-band oriented. The packet analysis gives where the basis is adaptive and is calculated using Shannon entropy of formula (6). Thus, for packet analysis we get output from . Essentially, after thresholding, the method measures energy at the output of filter banks as extracted live-iris pattern. The main idea behind the algorithm is that energy distribution is energy domain identifies a pattern and detecting conjunctival vascular pattern for developing liveness verification iris images. This shown in figure 3.

At step (22), the first and the second enrolled iris template mages are stored for further reference.

At step (24), a first and a second verification iris images of the human eye to be detected it captured as explained in step (12).

At step (26), the first and a second verification iris images are decomposed by using Daubechies wavelet based algorithm as explained in step (14).

At step (28), wavelet packets are used to extract information from activity band, which are decided using entropy formula. Activity area of more than 70% entropy change is further used for nodal analysis as explained in step (16).

At step (30), the first and a second verification iris images are subjected to sub-level decomposition for recognition of patterns as explained in step (18).

At step (32), the first and a second verification iris images are filtered for reconstruction these images to obtain seven high bands thereby obtaining energy distribution pattern and detecting conjunctival vascular pattern for developing liveness template as explained in step (20).

At step (34), the first and a second verification iris images are matched with the first and the second enrolled iris template mages, if there is a match the iris is genuine and live. If the match is not found the iris is not genuine or is not live.

Further, the energy distribution described above is the pattern being considered to analysis similarity between inter and intra class live-iris patterns, score based on the ‘Kullback-Leibler’ distance between two iris images is calculated. The ‘Kullback-Leibler’ distance is essentially relative entropy between two densities d1 and d2, and is given as below:
where, densities d1 and d2 represent two images under consideration. This is an attempt to characterize the live-iris pattern via marginal distributions of their wavelet sub-band coefficients.



Here, a is variance and ß is inversely proportional to the decreasing peak frequency. A good probability density function (PDF) approximation for the marginal density of coefficient at a particular sub-band is achieved by adaptively varying a and ß of GGD. Marginal distributions give better representation of live-iris pattern than the wavelet sub-band energies.
Using equation (14) and (15), closed from of the KLD is given:

The similarity measurement between two wavelet sub0bands can be computed very efficiently using the model parameters.

The overall distance between the images can be given as follows,

This is because of the scalable nature of the wavelet transform. The wavelet coefficient in different sub-bands are independent, and so to find overall score, individual KLDs are summed up. Here, j the sub-band level.

Also, the experimental calculations are done on the data set mentioned in section (2). Statistical independence of the live-iris patterns is directly proportional to the degrees of freedom used for doing similarity analysis. Moreover more degrees of freedom results in a more complex system. 1456 interclass and 169728 intra-class combinations were exploited. As, for the intra-class scores, as the data collection was performed over 5 months, consistency factor of the live-iris pattern is also studied. No environmental, emotional and psychological factors were considered.

The similarity score between the images is calculated using equation (16). This form of KLD is easy to implement and is found better than other close techniques like Bhattacharya coefficient [7]. A smaller similarity score indicates a better match. The in-band matching is done for individual bands, and then the scores are added and normalized. No further thresholding is implemented as the coefficients are already thresholded.

The method (100) ends at step (36).

Results are shown in Figure (4). The normalized rates are plotted on the y-axis and normalized similarity scores are plotted on x-axis. Inter class distribution is observed to be random, while intra class distribution is observed to be similar. Hardly any variation in the intra-class distribution is observed, thus confirming the consistency in the live-iris pattern of the 52 subjects over the period of 5 months. Distinct separation between the two classes is seen, thus proving accurate personal identification.

The live-iris pattern was observed to be `unique' for the collected data set. The pattern also showed good consistency within the same class when monitored over 5 months. By subjecting a known live-iris pattern for an individual, the iris recognition systems shall be robust to attacks. Beyond its possibility as a biometric alone or more promisingly in conjunction with iris, the method of unique liveness verification in iris scans can decrease the vulnerability of biometric devices to spoof attacks. The experiments were carried out for various infra red wavelengths and the best results were obtained for wavelength of 795 nm. Results of characterization of live-iris Pattern is shown in figure 4. The similarity score is normalized. The intra class can be seen very similar, and inter class can be seen distributed around 0.6, and hence very much random.

The foregoing descriptions of specific embodiments of the present invention have been presented for purposes of illustration and description. They are not intended to be exhaustive or to limit the present invention to the precise forms disclosed, and obviously many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the present invention and its practical application, and to thereby enable others skilled in the art to best utilize the present invention and various embodiments with various modifications as are suited to the particular use contemplated. It is understood that various omissions and substitutions of equivalents are contemplated as circumstances may suggest or render expedient, but such omissions and substitutions are intended to cover the application or implementation without departing from the spirit or scope of the claims of the present invention.