Vaccine Detection using Iris Extraction and Recognition

DOI : 10.17577/IJERTCONV10IS08024

Download Full-Text PDF Cite this Publication

Text Only Version

Vaccine Detection using Iris Extraction and Recognition

Mr. A. Joshva Issac

Assistant Professor

M.I.E.T. Engineering College Trichy, India

Vinitha Selvam

Computer Science and Engineering

M.I.E.T. Engineering College Trichy, India

Atchaya Arumugam

Computer Science and Engineering

M.I.E.T. Engineering College Trichy, India

Devika Ravichandran

Computer Science and Engineering

M.I.E.T. Engineering College Trichy, India

Abstract Vaccination is the most important things where the people are completely provided with the safe environment in this COVID 19 Situation. This vaccination helps to saves the people from the effect of this pandemic disease. The UNO and health organization belief that the vaccinated people among 80% are safe guarded from the COVID 19 virus. But many of them have not taken the vaccination where they go to public environment places like malls, theatres, temples, etc. From this people the vaccinated one can also be affected by the corona virus. In the existing system there is no measure to know the people who all are vaccinated or non- vaccinated. Thus in our proposed system iris based identification of vaccinated one and non-vaccinated one is enhanced. This system helps to avoid the non-vaccinated one in public place which reduce the spread of virus. Here, the system gives Decision Tree Classification algorithm with identification of the vaccinated people. Here a normal aadhar database system is created to know the people vaccinated with the matched feature of the iris extracted. This matching provides a classification of vaccinated and non- vaccinated people with their dose informations. Thus, the system completely helps for the safe and secure environment to reduce the spread of corona from non-vaccinated ones. This reduces the fear of COVID spreading in public places and allowing only vaccinated person. The efficiency of the process is high and time consumption of the recognition is completely low.

Keywords Ml,Haar,Rnn,


    Machine learning is a subfield of Artificial Intelligence. The goal of machine learning generally is to understand the structure of data and fit that data into models that can be understood and utilized by people. Although machine learning is a field within computer science, it differs from traditional computational approaches. In traditional computing, algorithms are sets of explicitly programmed instructions used by computers to calculate or problem solve. Machine learning algorithms instead allow for computers to train on data inputs and use statistical analysis in order to output values that fall within a specific range. Because of this, machine learning facilitates computers in building models from sample data in order to automate decision-making processes based on data inputs. Any technology user today has benefitted from machine learning. Facial recognition technology allows social media

    platforms to help users tag and share photos of friends. Recommendation engines, powered by machine learning, suggest what movies or television shows to watch next based on user preferences. Self-driving cars that rely on machine learning to navigate may soon be available to consumers. Machine learning is a continuously developing field. Because of this, there are some considerations to keep in mind as you work with machine learning methodologies, or analyze the impact of machine learning processes. In todays network world, the need to maintain the security of information or physical property is becoming both increasingly important and difficult. Most of the times, criminals have been taking advantage of a fundamental flaw in the conventional access control systems. The access control systems based on biometrics have a potential to overcome most of the deficiencies of current security systems and have been gaining importance in recent years. Comparison of various biometric traits shows that iris is very attractive biometric because of its uniqueness, stability, and non-intrusiveness. Number of problems required to be tackled in order to develop a successful iris recognition system, namely aliveness detection, iris segmentation, and feature extraction. However, iris is an annular part of an eye surrounded by other unwanted parts. Hence, imposes various challenges in accurate iris segmentation and feature extraction techniques to provide many opportunities for researchers in pursuing their research work in this area.


    Literature Review

    1. Dynamic Time Warping In Iris Biometric Recognition Process

      In general, iris reputation structures with linear normalization version forget about the Pupil Light Reflex. The pupil light reflex is accountable for adjusting mild depth reaches the retina and reasons nonlinear iris contraction and dilation actions that generates enormous variations among enrolled pix and testpix. This paper has proposed a way to lessen the influence of iris dynamics tested via way of means of decidability (d) and Equal Error Rate, received within side the evaluation among iris codes in specific states of dilation. The approach has used the

      Dynamic Time Warping (DTW) method to examine the Histograms of Gradients Oriented (HoG) vectors extracted from the iris texture. In this way, the maximum discriminated traits among testpix and the gallery were aligned and as compared, thinking about a nonlinear deformation of the iris tissue resulting from the PLR. The experimental results, the usage of dynamic pix, suggest the machine overall performance worsens while as compared to pix in specific states of contraction..

    2. A Robust Iris Segmentation Scheme Based On Improved U-Net

      Iris segmentation performs an vital function with inside the iris reputation system, and accurate segmentation of iris can lay a terrific basis for the follow-up 7 paintings of iris reputation and can substantially enhance the performance of iris reputation. We proposed 4 new viable community schemes and the first-class community version Fully Dilated Convolution combining U-Net (FD-UNet) is acquired via way of means of training and trying out at the equal datasets. FDUNet makes use of dilated convolution rather than unique convolution to extract extra international features, in order that the info of photos may be processed better. The proposed technique is examined with inside the near-infrared illumination iris datasets (CASIA-iris- interval-v4.0, ND-IRIS-0405) and the seen mild illumination iris dataset (UBIRIS.v2). The f1 ratings of our version at the CASIA-iris-interval-v4.0, ND-IRIS-0405 and UBIRIS.v2 datasets reached 97.36%, 96.74%, and 94.81%, respectively. The experimental outcomes display that our community version improves the accuracy and decreases the mistake rate, and plays properly on each near-infrared illumination and seen mild illumination iris datasets with suitable robustness.

    3. Assessing The Impact Of Corneal Refraction And Iris Tissue Non-Planarity On Iris Recognition

      In iris recognition algorithms, assumptions about eye morphology are made implicitly. When the eye look is focused on the iris texture, the cornea is thought to have little to no effect. The iris is supposed to be flat and the eye to be non-frontal. An orthographic projection will be used to create the image. If these assumptions are correct, If this is true, affine transformations can be employed to correct nonfrontally shifted data. The rubber sheet model and posed photographs to a frontal view be used to precisely normalize the iris annulus to a rectangular shape image. This research looks into how iris recogntion works. When the first two 8 assumptions are broken, the result degrades. Making use of a. A data collection changing in size is constructed using a computer render able eye model. The cornea's presence, as well as a parameterized non-planarity is formed on the basis of Iris-related. Matching rankings are created the use of a industrial matcher. When evaluating the relative effect of every assumption violation, itis found that iris nonplanarity gives a greater significant hassle than corneal refractive distortion in regards to iris popularity accuracy in non-frontal images.

    4. Robust Iris Presentation Attack Detection Fusing

      2d And 3d Information

      Diversity and unpredictability of artifacts potentially presented to an iris sensor calls for presentation attack

      detection methods that are agnostic to specificity of presentation attack instruments. This article proposes a method that combines twodimensional and three- dimensional properties of the observed iris to address the problem of spoof detection in case when some properties of artifacts are unknown. The 2D (textural) iris features are extracted by a state-of-the-art method employing Binary Statistical Image Features (BSIF) and an ensemble of classifiers is used to deliver 2D modality-related decision. The 3D (shape) iris features are reconstructed by a photometric stereo method from only two images captured under near-infrared illumination placed at two different angles, as in many current commercial iris recognition sensors. The map of normal vectors is used to assess the convexity of the observed iris surface. The combination of these two approaches has been applied to detect whether a subject is wearing a textured contact lens to disguise their identity. Extensive experiments with NDCLD'15 dataset and a newly collected NDIris3D dataset show that the proposed method is highly robust under various open-set testing scenarios, and that it outperforms all 9 available open-source iris PAD methods tested in identical scenarios. The source code and the newly prepared benchmark are made available along with this article.

    5. Ensemble Of Multi-View Learning Classifiers For Cross-Domain Iris Presentation Attack Detection

    The adoption of large-scale iris recognition systems around the world has brought to light the importance of detecting presentation attack images (textured contact lenses and printouts). This paper presents a new approach in iris presentation attack detection (PAD) by exploring combinations of convolutional neural networks (CNNs) and transformed input spaces through binarized statistical image features (BSIFs). Our method combines lightweight CNNs to classify multiple BSIF views of the input image. Following explorations on complementary input spaces leading to more discriminative features to detect presentation attacks, we also propose an algorithm to select the best (and most discriminative) predictors for the task at hand. An ensemble of predictors makes use of their expected individual performances to aggregate their results into a final prediction. Results show that this technique improves on the current state of the art in iris PAD, outperforming the winner of LivDet-Iris 2017 competition both for intra- and cross-dataset scenarios, and illustrating the very difficult nature of the cross-dataset scenario.


    In the existing system, a deep network based iris detection system is produced by the author. Despite the promise of recent deep neural networks to provide more accurate and efficient iris recognition compared to traditional techniques, there are vital properties of the classic Iris Code which are almost unable to be achieved with current deep iris networks: the compactness of model and the small number of computing operations (FLOPs). This paper casts the iris network design process as a constrained optimization problem which takes model size and computation into account as learning criteria. On one hand, this allows us to fully automate the network design process

    to search for the optimal iris network architecture with the highest recognition accuracy confined to the computation and model compactness constraints. On the other hand, it allows us to investigate the optimality of the classic Iris Code and recent deep iris networks. It also enables us to learn an optimal iris network and demonstrate state-of-the- art performance with less computation and memory requirements.



    In our proposed system, a RNN based iris matching vaccination detection is done. Here the identification of the people who are vaccinated or not is our main focus. The system makes a identification of Iris of a particular person with computer vision processing. With the detection of the iris Aadhar database connectivity provided which gets the details of the vaccinated and nonvaccinated person. Thus the person with vaccination will be detected easily from the proposed connectivity approach. The main objective of the system is to classify the vaccinated and the non-vaccinated people from the public places. This classification will be done with the identification of the vaccinated people on the connectivity of the database approach on iris categorization. Here webcam based iris detection will be done where the feature matching HAAR cascades employs a wide role on extraction of the iris works. Besides that, flexibility in terms of accepting the appropriate input image is seeking for improvement. The ready solution is through the effort made on both hardware and software, purposely to eliminate the unwanted features and highlighting the important details of the image captured. Here the system will be connected and detected with the vaccinated details stored in the aadhar details server which shows out the classification of the vaccinated person and the non- vaccinated person in the public places. Thus this mainly shows not allowing the nonvaccinated person in public on detection using the developed project work.

      1. Video Acquisition

        Video acquisition is the process by which video surveillance data input given to the system. The video acquisition process will gather the video data and convert it as a frame conversion. Frame conversion is the one where the information will be converted as frames so that the identification will be easier as they will be as images. The

        complete identification will be done with the running live video with different co-ordinate points. Video acquisition those are relevant with regards to the concept of format- agnostic media production. The format-agnostic scene representation can contain ultra-high resolution panoramic video and several views from broadcast cameras in the best case. Several approaches are presented showing the capability for seamless merging of various video sources. Regarding higher frame rate, it is shown that an increase of frame rate is not just a matter of technological progress per se, but it is an essential property to increase video quality and immersiveness, especially in the case of significantly increased spatial resolution. Acquisition with higher dynamic range can offer increased video quality by introducing tone mapping approaches during media production.

      2. FACE DETECTION WITH IRIS EXTRACTION The face detection algorithm looks for specific Haar features of a human face. When one of these features is found, the algorithm allows the face candidate to pass to the next stage of detection. A face candidate is a rectangular section of the original image called a sub- window. Generally these subwindows have a fixed size (typically 24×24 pixels). This sub-window is often scaled in order to obtain a variety of different size faces. The algorithm scans the entire image with this window and denotes each respective section a face candidate. The algorithm uses an integral image in order to process Haar features of a face candidate in constant time. It uses a cascade of stages which is used to eliminate non-face candidates quickly. Each stage consists of many different Haar features. Each feature is classified by a Haar feature classifier. The Haar feature classifiers generate an output which can then be provided to the stage comparator. The stage comparator sums the outputs of the Haar feature classifiers and compares this value with a stage threshold to determine if the stage should be passed. If all stages are passed the face candidate is concluded to be a face.


        In this module the iris will kept matched with the trained Aadhar database. In the database the iris co-ordinate points will be get stored where the system will exactly classify the person name then the iris feature points get matched and extract the Aadhar card number from the database. This helps the system for the further identification of the system which shows out further classification system. The extracted iris part will be made with two executions training and the testing phase: the training part will be done with complete analysis of the feature trained with Aadhar server. Thus with the trained pattern the extraction over the identified feature will again matched with the server on extraction generation.


        The Aadhar database connectivity system will be given with a different pattern of extraction with iris features. The iris feature extraction will be the matched with the

        database of the Aadhar system. The system correctly extracted trained iris pattern with the Aadhar number in it and shows out.

      5. Vaccination Classification

    The vaccination classification system is generated with a high level of RNN processing system. The RNN classification is done through the processing implementation which is carried out. The sub phase will make a detection over the identified features of the iris extracted


      This paper has provided a groundbreaking and modern approach for vaccination detection of COVID-19. It also illustrates the mechanism of the proposed COVID-19 detection system. The reason behind these in-efficient preliminary results is that due to time constraints, the collected data set is comparatively small and lacks control group data on healthy subjects. The Haar classifier is designed using a pipelined scheme, and the triple classifier which three single classifiers processed in parallel is adopted to accelerate the processing speed of the face detection system. The exact iris detection with a matching shows out an exact clarification system of the trained feature pattern. Thus a efficient safe guarding of people from non-vaccinated people will be done at 99% accuracy system on each algorithm execution.


[1] Bhanu.B and Kumar.A Deep Learning for Biometrics.

Springer, 2017.

[2] Boyd.S and Vandenberghe.L Convex Optimization. Stanford University, 2004.

[3] Bubeck.S Convex optimization: Algorithms and complexity.

Foundations and Trends in Machine Learning, 2015.

[4] Chinese Academy of Sciences Institute of Automation. CASIA iris image database, Aug 2017.

[5] Daugman.J How iris recognition works? IEEE Transactions on Circuits and Systems for Video Technology, 2004.

[6] Daugman.J New methods in iris recognition. IEEE Transactions on Systems, Man and Cybernetics, 2007.

[7] Daugman.J Information theory and the iriscode. IEEE Transactions on Information Forensics and Security, Feb 2016.

[8] Daugman.J andDowning.C Searching for doppelgangers: assessing the universality of the iriscode impostors distribution. IET Biometrics, 2016.

[9] Franceschi.L, Frasconi.P, Salzo.S, Grazzi.R, and Pontil.M Bilevel programming for hyperparameter optimization and metalearning. In International Conference on Machine Learning (ICML), 2018.

[10] Gangwar.Aand Joshi.ADeepirisnet: Deep iris representation with applications in iris recognition and cross-sensor iris recognition. In IEEE International Conference on Image Processing (ICIP),Sep 2016.