INTRODUCTION:

In Biometrics, Iris recognition is one of the definitive techniques in human identification using biological, physiological traits or habits.

To further extend the reliability of the human-in-the-loop integrating a quality measure for enrolled images would be favorable for biometric system [1]. The iris recognition algorithm utilizes the rapid attenuation created by binomial combination to avoid the suffering of false matches for very large data bases [2]. To explore the use of macro features of the upper surface of RGB images for matching and recovering. The evaluation is resolved by the number of times the iris containing the macro features which is reclaimed from gallery space [3]. Using Daugman algorithm, each iris pattern was encoded into 2048 bits of face sequence. A basic logic of shape description teaches that weak controls must be used for strong data’s and strong controls to be used for weak data’s which make the pupillary boundary fit in an active contour [4].

The degraded images which are acquired from less constrained condition can be handled using the method of segmentation. Contributions are added to make the iris segmentation to be successful in challenging condition 1) The sclera is considered to be the easily detectable part of the eye 2) A new type of feature is proposed to measure the sclera proportion. 3) The entire procedure is executed to make it suitable for real time application. This paper comprises this ambitious condition that guide to degradable image data which can be challenging for segmentation [5]. In iris recognition crypts and anti-crypts which are the two visible features are explored when compared with finger prints. Iris physical features do not change during adulthood and thus becomes the most beneficial biometric modes for authentication and recognition. Crypts are visible in the iris surface areas which relates to pigmentation. Anti-crypts are found in the thick anterior layer. The results are more accurate, since the crypts and anti-crypts are visible and can be understood easily [6]. Multi scale pyramid architecture is used by the system to detect the features of candidate before they are analyzed and optimized by heuristic based method. A two-stage matcher is used to measure the dissimilarity between irises [7]. Iris recognition and palmprint recognition have an effective feature representation called as ordinal measures. Variants with parameter settings such as location, scale, orientation etc. Image analysis has ordinal measures as a common concept. For ordinal feature selection, a novel optimization formulation has been proposed with applications that are successful to both iris recognition and palmprint recognition. For linear programming (LP) problems the ordinal feature selection has been formulated so that for large scale attribute pool and training database an effective solution can be obtained. The LP feature selection has been a success for iris recognition and palmprint recognition [8]. Segmentation based background removal (SBR) and triangular shapes DCT (tri DCT) are used for enhanced iris recognition (IR) system. The desired Region of Interest (ROI) has been received successfully by SBR technique. Optimal count of feature subset can be extricated by the help of triDCT and BPSO. Iris Mask (IM) is created using fixed radius length using a static approach [9]. Enhancement of iris recognition is done by a distinctive combination of edge detection plus DCT based feature extraction. Circular sector and triangular shaped DCT feature extraction techniques are the two methods which are proposed for iris enhancement. BPSO algorithm is used for obtaining optimal feature. A considerable reduction subset in the number of features has been discovered with circular sector and triangular shaped DCT extraction and BPSO applications [10]. To increase the overall accuracy for iris recognition system, the performance of segmentation and normalization place an important role [11]. The process of isolating the object from the features is called segmentation [9]. Good segmentation is provided by the CASIA database. The sclera and boundaries were clearly differentiated e from the eye images that were chose for iris recognition research. The system for detecting eyelash has been applied for the CASIA database which isolates the eyelashes that occurs within the iris region. This system has been proved to be successful in isolating the eyelash [11]. Iris segmentation, quality enhancement match score fusion and indexing algorithms are used to increase the accuracy and the speed of iris recognition. The modified Mumford-Shah function is performed to segment the iris image which is non-ideal. This is a curve evolution approach. A locally enhanced region is selected by a support vector machine based learning algorithm from globally enhanced image. These regions with good quality are combined to form a single high-quality iris image. The topological matching scores are combined by an intelligent fusion algorithm to enhance the performance of iris recognition. This also reduces the false rejection rate. Fast and accurate iris identification is done by an indexing algorithm. The performance of iris verification and identification has been improved [12]. A hardware-software hybrid method is used to improve the stand-off distance in an iris recognition system. To extend the depth of field, an optimized wave front coding technique is used. To restore the blurred image on the software side local patch-based super- resolution method is implemented and to bring the image to its clear version. Capture volume of a conventional iris recognition system can be improved by three times and the system’s high resolution rate can be maintained. The quality of the captured iris image can be enhanced effectively by the proposed CPSR, through the PSF invariant property application for various defocus positions [13]. To address degradation while matching iris images, a domain adaptation frame work has been developed. Using Markov random field model, a new algorithm has been introduced to improve cross-domain iris recognition significantly. From three publicly available databases, experimental results are presented. Cross sensor and cross-spectral iris matching are also achieved. The three databases are poly u cross-spectral iris image database, IIITD, CLI and UND database [14]. In pre-developed iris recognition system, optimum path forest is used as a classifier. The optimum path forest is a graph based framework for pattern recognition. Experiment with different scale database are performed to access the performance of OPS for the iris recognition tasks and these results are compared with Hamming distance spaced classifier and Bayesian classifier [15][16]. The performance of optimization of some framework is number of hidden layers and knowledge for the training phase. By using firefly algorithm better results were obtained. These databases are used for effectiveness and efficiency of the method used by an objective functioning of the recognition of the error. This was used for pattern recognition based on the biometrics of iris [17][18]. Negative database is the best skill for privacy preservation. This method has highly recognition performance on the typical database. The recognition of the result was 99.13% [19][20][21]. The noise-signal-level information [22][23] is used to mitigate the impact of noise and degradation for minimum iris recognition systems [24][25]. This is purely based on the low rank approximation (LRA). Multiple noise [26][27] of the eye is captured in this approximation [28][29].

OVERVIEW:

This paper aims in reducing the noise level from the extracted crypts in order to avoid false matching. The proposed system consists of five modules; the first step is Pre-Processing where Gaussian filter is used to reduce the image pixels to get a clear image. Triangular filter is used in filtering process and image is enhanced using gradient techniques. Multilevel image threshold using Otsu’s method is used to obtain the segmented image in the segmentation process. This step is followed by feature extraction where area, bounding box, centroid and convex area are extracted. The final process is similarity analysis which has the output in graphical representation and compares the existing and proposed system.

GRAY LEVEL FEATURES:

The existing system is crypt features extraction from iris data sets. This system is based on gray level features. Gray Level Features are extracted. The existing system automatically acquires the biometric data in numerical format (Iris Images) by using a set of correctly located sensors. These are considering camera as a high-quality sensor. Iris Images are typically color images that are processed to gray scale images. To identify an individual by comparing the feature obtained from the feature extraction algorithm with the previously stored feature by producing a similarity score. Take 24-bit BMP color image as input. Then convert it to 8-bit Gray Scale image by following this algorithm. This 8-bit Gray Scale image is the output of the algorithm. In this algorithm, read the red, blue and green value of pixel and after that formulation, three different values are converted into gray value. By Statistical Analysis, IRIS PATTERN of an individual can be generated which can be strongly used for Pattern Recognition or over all Human Recognition. The iris texture recognition for the purpose of human identification is by means of statistical analysis of gray-level distribution. A lot of studies have been intended at extracting iris features that are unique to every individual. While many have been successful, most requires complex filtering and processing.

The literature survey describes crypt features extraction from iris data sets. In [5] Ordinal Feature Selection for Iris and Palm Print Recognition is a novel optimization formulation for ordinal feature selection with attractive applications to both iris and palmprint recognition. The feature selection aims to achieve an accurate and sparse representation of ordinal measures. The optimization subjects to a number of linear inequality constraints, which need that all intra and interclass matching pairs are well separated with a large margin. The use of macro-features that is observable on the anterior surface of RGB images of the iris for matching and retrieval. Moles, freckles, nevi, melanoma, etc. are structures to which these macro features correspond and may not be present in all iris images. The use features extracted by the Scale-Invariant Feature Transform (SIFT) to represent and match macro-features. In segmentation method, the sclera is most easily distinguishable part of the eye in degraded images, a new type of feature that measures the proportion of sclera in each direction and is fundamental segmenting the iris, and to run the entire procedure in deterministically linear time in respect to the size of the image, making the procedure suitable for real-time applications.

BLOCK DIAGRAM:

Original image	Pre-Processing	Filtering

Similarity Analysis	Feature extraction	Segmentation

METHODOLOGY:

A. PRE-PROCESSING:

The operative searches for the circular path anywhere is maximum change in pixel values, by varying the radius and center x and y position of the circular contour. The operator is practical iteratively with the amount of smoothing gradually reduced in order to attain precise localization. In a similar manner, eyelids are localized with the path of contour integration changed from rounded to an arc. The essential can be seen as a variation of the Hough transform, since it uses the image’s first derivatives and performs a search to find geometric parameters. As it works with raw derived information, it does not suffer from the thresholding problems of the Hough transform. Gaussian filter has Gaussian functions as impulse response. Gaussian function is represented in equation.1.Convolution matrix is calculated and convolved with the data. The focal element has maximum weights and the remaining elements possess smaller weights. However, the algorithm can fail where there is noise in the eye image, such as from reflections, since it works only on a local scale. In order to encode features, a system decomposes the iris region by application of Laplacian of Gaussian filters to the iris region image [6].

ae^-(x-b) ^2 (1)

F(x) = ---------------------------------

2*c^2 (1)

Where a, b and c are arbitrary constants.

FILTERING:

TRIANGULAR FILTER:

The triangular Filter techniques based ciliary features are extracted from iris data sets. The ciliary features are extracted by various techniques are Gradient level techniques are used. An original approach to person recognition based on iris patterns, spectacles contact lens wearing persons and diseased eyes having indoor and outdoor conditions. A challenge-response method is used for eye aliveness test that puts off artificial sources as of incoming the iris database. With 84 statistical iris features that are extracted from an individual, the proposed algorithm can work. Space and time density of the projected approach is lesser than the existing methods. A new algorithm for identification of persons using human iris patterns that works under indoor and outdoor conditions, with eyewear or diseased eyes. The challenge response test is used for checking eye aliveness, which ensures detection of counterfeit irises, and is compared with data enrolled from the iris-encrypted file. In the biometrics, such as the palm, retina, gait, face and fingerprints, the characteristic of the iris is stable in a person’s lifetime. Iris patterns are wildly distributed and well suited for recognizing persons during their lifetime with a single conscription.

Image Enhancement using Gradient Techniques:

A method for extract an iris from an image is obtainable [7]. Edges of an iris are detected in an image. Texture from the image is acquired. The edges and texture are combined to generate inner boundary and an outer boundary of the iris. Normalized histogram equalized values are shown in equation.2. To improve the accuracy of the iris boundary detection, a method to select between ellipse and circle models is provided. Finally, a dome model is used to determine mask images and remove eyelid occlusions in unwrapped images.

Number of pixels with intensity n

V= --------------------------------------------------------

Total number of pixels (2)

Where n=0, 1…, L-1.

L represents number of possible intensity values

V represents the normalized histogram of image

The histogram equalized image is defined as,

Image _{(i, j)}= floor((L-1)∑ _n=0^f(I,
j)V) (3)

Where floor rounds nearby negative infinity values.

C. SEGMENTATION:

The Multilevel image threshold uses Otsu method to obtain segmented image [8]. A principle for maximizing the between-class variance of pixel intensity to perform picture thresholding. On the other hand, Otsu’s method for image segmentation is very time-consuming because of the inefficient formulation of the among -class variance. A faster version of Otsu’s method is potential for improving the efficiency of computation for the optimal thresholds of image segmentation.

Then the new criterion is designed to efficiently find the optimal threshold. This procedure yields the same set of thresholds as the original method. In addition, the modified between-class variance can be pre-computed and store in a look-up table. Our analysis of the new criterion clearly shows that it takes less computation to compute both the cumulative probability and the mean of a class, and that determining the modified between-class variance by access a look-up table is earlier than that by performing mathematical arithmetic operations. For example, the experimental results of a five-level threshold selection. The method can reduce down the processing time from more than one hour by the conventional Otsu’s method.

D. Feature extraction:

The features are extracted by investigative the structural information controlled in the distribution of minutiae points using Delaunay triangulation. The goal of an iris recognition system is to extract features representing the textural information present in the iris and thus authenticate (verify or identify) an individual, on the basis of such features. The main modules of such a system are segmentation, enhancement, feature extraction and matching. During enrollment, the features are extracted and stored in the database as templates. The authentication phase encompasses image preprocessing followed by feature extraction for a given iris image. This feature set is then compared with the templates in the database in order to perform identification or verification of an individual’s identity.

E. CILIARY FEATURES:

Feature extraction a type of dimensionality decrease that efficiently represents interesting parts of an image as a compact feature vector. This move toward is useful once image sizes are large and a reduced feature image is required to quickly complete responsibilities such as image matching and recovery. The detection and extraction, as well as matching are frequently shared to solve common computer vision problems such as object detection and recognition, content-based image retrieval, face detection and recognition, and texture classification.

F. IMAGE GRADIENT:

The image gradient is a directional change in the intensity or color in an image. Image gradients may be used to extract information from images. Iris recognition systems are mainly dependent on the presentation of iris localization processing. Ranking after localization requires normalization, feature extraction and matching. These steps are based on the accuracy and efficiency of localization of iris in human eye images. the inner boundary of iris is calculated by finding the pupil center and radius using methods, selected region is adaptively binarized and centroid of the region utilize for obtaining pupil parameters. Edges are process to detect radius and center of pupil during the following method. Used for outer iris boundary, a band is calculated within which iris outer boundary lines.

SIMULATION OUTPUTS:

a. Input image b.Grey type image

c. Gaussian filtered smoothened image

d. Triangulared enhanced image e. Enhanced image

f. Multilevel threshold image g.Ciliary feature

h. Ciliary binary image i. Ciliary image gradient

j. Ciliary image direction

k. CILIARY Feature Dissimilarity

CONCLUSION:

A system designed for performing iris recognition may consist of a processor which controls an illumination control circuit and a camera to obtain images at several predetermined sizes of the pupil. The aim of this project is to implement a working triangular filtering technique of biometric system for human authentication recognition application development and methods used for iris recognition, and to test these methods on a database.

REFERENCE:

1. Jianxu Chen, Feng Shen, Dannyz-Chen,” Iris recognition on based on human interpretable features “, pp.1-10, No.1, 2016.

2. J.Daugman, ” How Iris Recognition Works”, vol:14, pp.21-30, issue 1, 2004.

3. M.Sunder and A.Ross, “Iris Image retrival based on macro features” pp.1318-1321, 2010.

4. John Daugman, “Probing the uniqueness and randomness of Iris codes”, vol:94, pp.1927-1935, No.11, 2006.

5. Hugo Proenca, “On Segmentation of degradable images acquired in the visible wavelength”, vol:32, No.8, 2010.

6. Feng Shen, Patrick J.Hynn, “Iris matching by crypts and anti-crypts, pp.208-213, 2012.

7. Feng Shen, Patrick J.Hynn, “Multi-scale detection and shape-based matching, pp.977-983, 2014.

8. Z sun, L. Wang and T. Tan, “Ordinal features selection for iris and palmprint recognition”, vol:23, pp.3922-3934, No.9, 2014.

9. Amrinder Singh Randawa, Akula Radheya and Manikatan, “Segmentation based background removal technology for enhanced iris recognition”, 2015.

10. Abhiram M.H, Chetan Sadhu, K.Manikatan, S.Ramachandran, “Novel Dct based feature extraction for enhanced iris recognition”, 2012.

11. Surjeet Singh, Kulkir Singh, “Segmentation technique for iris recognition system”, vol:4, 2014.

12. Mayank Vasta, Richa Singh and Afzel Noore, “Improving iris recognition performance using segmentation, quality enhancement, match score fusion and indexing”, vol:38, No.4, 2008.

13. Sheng-Hsun Hsieh, Yung-Hui Li, Chuag-Hao Tien and Chin-Chen Choug, “Extending the capture volume of a iris recognition system using wave front coding and super-resolution”, vol:46, 2016.

14. Pattabhiramaiah, Nalla and Ajay Kumar, “Toward more accurate iris recognition using cross-spectral matching”, vol:26, No.12, 2017.

15. Luis C.S. Afonso, JoooP. Papa and Aparacidon.Moroma, Ahmad Poursabari and Svotiono N. Yanushkevich, ” Fast large scale iris database classification with optimum-path forest technique. A case study”, 2012.

16. Daniela Savchez, Patricia Melin, Juan Carpio, Hector Puga, “A firefly algorithm for modular granular neural networks optimization applied to iris recognition”, 2016.

17. Dongdong Zhao, Wenjian Luo, Ran Liu and Lihua Yue, “Negative iris recognition”, 2015.

18. Yang Hu, Konstantinos Sirlantizis and Gareth Howells, “Signal-level information fusion for less constrained iris recognition using sparse-error low rank matrix factorization”, vol:11, No.7, 2016.

19. Maflin Shaby, S. and Vimala Juliet, A. “Analysis and Optimization of Sensitivity of a MEMS Piezoresistive Pressure Sensor”, Advanced Materials Research, ISSN1662-8985, Vol. 548, pp. 652-656, 2012.

20. Maflin Shaby, S. and Vimala Juliet, A. “Analysis of Sensitivity and Linearity of SiGe MEMS Piezoresistive Pressure Sensor”, Applied Mechanics and Materials, ISSN 1662-7482, Vol.241-244 pp.1024-1027, 2013.

21. Maflin Shaby, S. and Vimala Juliet, A. “Performance Analysis and Validation of sensitivity of Piezoresistive MEMS pressure sensor’, Proc. of International conference IEEE Recent Advances in Intelligent Computational Systems (RAICS2011), organized by IEEE, Thiruvananthapuram, pp. 692-695, ISBN 978-1-4244-9477-4, 2011.

22. T.Gomathi, :” Performance analysis of parallel co-operative spectrum sensing using scheduling algorithm in CRNS”, International Journal of Applied Engineering Research ISSN 0973-4562, Volume 10, Number 2 (2015) pp. 2733-2747, 2015.

23. Ms.T.Gomathi, Dr.S.Maflin Shaby “Capacitive accelerometer for Micro Electromechanical Applications :A Review "has been presented in the International IEEE Conference ICCICCT-2016, 2016

24. Bestley Joe S, Ms.T.Gomathi, Dr.S.Maflin Shaby “efficient low wind energy harvesters for wireless sensor networks”, Journal of Materials and Environmental Sciences, ISSN : 2028-2508, vol 9, pp.622-630, 2018.

25. Bestley Joe S, Ms.T.Gomathi, Dr.S.Maflin Shaby and Pandian.R, “Driving Aid in Cloudy Conditions using Contrast Restoration”, “International Conference on Electrical, Electronics, Computers, Communication, Mechanical and Computing (EECCMC)”, Jan 2018.

26. Bestley Joe, S., Gomathi, T., Maflin Shaby, S., Pandian, R.,” Self-assistance devices for asthma patients using android application “, Research Journal of Pharmacy and Technology, 2018

27. T.Gomathi, :” Compression and Segmentation of JPEG Images Using DWT” Research Journal of Pharmaceutical Biological Chemical Sciences (ISSN: 0975-8585), Volume 6, issue 2(March-April), 2015.

28. Maflin Shaby, S., Gomathi, T,” A wireless surveillance robot with motion detection and metal sensors for military purpose with live video transmission,” Research Journal of Pharmaceutical, Biological and Chemical Sciences, V olume: 7 Issue: 6 Pages: 1609-1613, 2016

29. Ms.T.Gomathi, Dr.S.Maflin Shaby,” Design of Fire Detection sensor and monitoring system, Research Journal of Pharmaceutical Biological Chemical Sciences (ISSN: 0975-8585), 2016

30. Ms.T.Gomathi, Dr.S.Maflin Shaby and B. Priyadharshini, “Instill Dental Antennas for Minimally Interfering Bio-medical Devices, Springer publication – ICCVBIC, pp 1049-1060, 2017.

Received on 13.03.2019 Modified on 25.04.2019

Research J. Pharm. and Tech 2019; 12(8):3759-3763.

DOI: 10.5958/0974-360X.2019.00644.9