Attention Makers


Smart Autonomous Multipurpose Biometric Security System

MAKERS: Sushmita Goswami

Purpose of our(including me and my project member Suchismita Goswami) project is to make an autonomous biometric security system which can be used to prevent attacks in various sensitive areas like airport,banks,passport system,prison etc. Technologies used Cognitive computing, Image Processing,Internet Of Things,Data mining,Android etc. # The problem that this system is going to solve is described as follows- 1. Airport :- Today?s passenger and cabin baggage security screening works, but at great cost to authorities, to the airline industry, and to passengers. Given the predicted growth in air travel, continuously evolving security threats, and passengers being increasingly dissatisfied with queues and intrusive measures.Moreover carry-on-items of the passengers are checked through dual-energy X-ray system, but the checking of the identity of travellers is done by checking a photo ID, such as a driver's license,passport etc.Simply taking a look at a photo ID isn't enough.In the proposed system travellers do not need to carry any smart card with him/her and the entire system would work without intervention of any security guard.This system can seeks to deliver strengthened security,Increased operational efficiency,Improved passenger experience etc. 2.Banks:- In banking system usually the account number or pin number and the signature is used for enabling security measures.The fraudster then uses this information to carry out fraudulent transactions in countries where chip and PIN technology is not supported.The fraudster may also use this information in transactions where the card doesn?t have to be physically seen by the retailer or merchant. For example, when shopping online.Proposed system provides another security layer in addition to the existing measures to prevent this spoofing and skimming attacks.The use of biometrics in banking can be used because of its potential as a predominant method of identifying themselves to access banking services such as branch banking, online banking, ATM networks, and mobile banking. 3. Passport:- The duplicate passport problem can also be resolved through this proposed system. 4.Terrorist Attack:- Now-a-days it is very much common that prisoners sometimes fled away from prison.If some facial changes takes place due to plastic surgery it become very difficult to identify though that person is suspiciously caught again for his behaviour. 5.This system can be used to screen people at our borders. 6.Patient Identification:- Misidentification of patients is a direct cause of preventable medication errors, transfusion errors, and testing errors.This system with biometrics can help to reduce the challenges of patient data matching and data sharing across disparate health networks. 7.Biometric Car:-The driver?s own biometric empowered vehicle might detect driving impairment, pull off the road and shut down the ignition until a replacement driver takes over.Proposed system holds promise for tighter anti-theft and anti-DUI enforcement measures.Accident investigations and crash prevention may become easier as a result of this new innovation. 8.Proposed system can be used to provide biometric security to the homeowners. ## Solution and implementation:- 1. Proposed system would work in Airport in following way :- It is implemented by analyzing the data obtained by retinal scan of an individual. At first it will scan the retinal images. Now the system will start processing, in the meantime another camera will automatically start and capture a facial image of the person. The system will extract the several features from the retinal image which will be used to develop the template. An algorithm to identify the exact location of the optic disc has been implemented. This algorithm is a new approach that gives the coordinate of the optic disk. Another algorithm for identification of the position of macula is introduced. Considering the fact that the image of retinal scan can be taken from various angles, the differences in the respective images can be overcome using a relative distance between macula and the optic disk. This is a new approach used in this algorithm that has not been previously utilized elsewhere. Now we record the bifurcation angles at each bifurcation points as these points are stable. Hence our algorithm effectively removes the crossover points and only identifies the stable bifurcation points. Another significant part of our algorithm is the identification of retinal diseases like glaucoma and Proliferative Diabetic Retinopathy. During glaucoma, the width of nerves vary and generation of fragile vessels around the optic disc occurs. Our prototype successfully identifies these fragile vessels around the optic disc and discards these vessels. As changes occur in the width as well, so width of retinal vessels is not considered as a biometric feature for generation of template. These characteristics are used to make the 96 bit template, and then a client program written in java will take this template as input and matches it with existing template in the database. If it matches then it will not send any information to the server in the security room, if it does not match then it will send this information and the image captured by the secondary camera to the server computer in the security room. Upon receiving this information to server computer, a pop up will be displayed and the picture of intruder will be shown and additionally one alarm will ring. Now in the server computer a chat window will be opened which will be connected through Wifi to the mobile of the security guard so that officers at security room can talk with the security guard. This has been implemented in android. For further analysis, the details of the intruder is stored in another database. This database is made public to all the peer organizations. This database helps us to keep a track of the places that the intruder has visited and count the number the intruder has approached. From this information through data mining technology we can analyze the probable places that the intruder may attack in future so that prior action can be taken. 2.Proposed system would work in smart car in following way:- At the time of purchase following procedure will be followed:- 1. Enroll : - Retinal scan 2. Feature Extraction (exact location of optic disc, no of bifurcation points ,bifurcation angles etc.) 3. Generation of unique 96 bit template using above extracted features. 4. finally it will store the template in the database of that system embedded in the car. This can be used as biometric lock for the car which will hold promise for tighter anti-theft and anti-DUI enforcement measures. This system will work in the same way with few modification in other applications too. ### Potential Advantage of this proposed of system:- 1. Current biometric authentication systems(retinal) are unable to detect and overcome the changes in the biometric feature due to diseases. This was the primary challenge for developing a biometric authentication and it has been successfully overcome in this system. 2. Images captured from different angles may lead to even authentication failure . In the proposed system this problem has been resolved as the relative features of different biometric parameters of retinal image will be considered. 3. .Existing systems are time consuming process, takes minimum 15 seconds time to adjust camera with human eye. It?s not accepted. In our proposed system it is approximately taking around 7-8 seconds. 4.Moreover another newest application of this system is ability to detect PDR(Proliferative Diabetic Retinopathy) at its earlier stage.Thus it is bringing an enormous progress in medical field also. ###Following are the challenges faced and solved during development :- Changes made during Glaucoma has been uniquely identified and discarded. Detection of Proliferative Diabetic Retinopathy. Hence this system can work for the identification of the person having any retinal diseases which remains a limitation of the current retinal scanning system.This prototype has been implemented and checked with sample retinal images with glaucoma disease and the system works with 99% accuracy. Moreover the software used like java,matlab,hadoop all are open source. Only cost includes the cost of retinal scanner.Hence this system is highly cost effective.

Home Previous Next

Vote Share Comment