Abstracts of Undergraduates Dissertations-UWU

Permanent URI for this community

https://erepo.lib.uwu.ac.lk/handle/123456789/249

Browse

Now showing 1 - 20 of 39

Artificial Neural Network Based Signature Recognition and Verification System
(Uva Wellassa University of Sri Lanka, 2013) Prabodith, N.P.C.
The signature of a person is an important biometric attribute of a human being which can be used to authenticate human identity. In present day signature use in many transactions in day today life and we can see some people are trying to miss use signature for achieve their narrow goals. So goal of this project is provide strong way system recognize and verify hand written signature. Signatures are intrapersonal biometric attribute and it differs from people to people. Even collection of signature of one person also differs from each other. But when we consider that collection of signatures, there is certain pattern which follows all signatures. So Signature recognition is such kind of pattern recognition and human signatures can be handled as an image and recognized using computer vision and neural network techniques. With modern computers, there is need to develop fast algorithms for signature recognition. There are various approaches to signature recognition with a lot of scope of research. In here, off-line signature recognition & verification using neural network is proposed, where the signature is captured and presented to the user in an image format. Signatures are verified based on parameters extracted from the signature that using various image processing techniques. This System accepted image of image and generate single output. That output can be single digit or pattern which containing binary values. Digital Image Processing and Artificial Neural Networks techniques are main techniques used for implement system. According to the system DIP and ANN are two main parts in development side. When system accepts signature by image format and then its process using DIP. System gathered some unique features (Details) form Image and which use as input in ANN. In here for both Image processing and Artificial Neural Networks programming language based on C#.net with the help of Aforge. net computer vision library. Customized classes of Aforge. net Imaging library used for image preprocessing activities while all feature extraction activities algorithms implemented by using C#.net language. Aforge Neuro library used for ANN programming. Those two parts combined together and works as complete system.
Artificial Vision System for The Blind Using Ultra Sonic Wave Transmission Reception Techniques
(Uva Wellassa University of Sri Lanka, 2013) Hisham, M.H.M.
Various devices and systems have been identified for the usage of blind people which can helps to travel safely and efficiently. These systems are based on the ultra sonic sound transmission and reception techniques to determine the distance between blind people and objects along the path of the blind people while they are walking. But most of the time these systems are having limitations and most blind people do not prefer to use the existing systems, because of the sound which is produced by the device. The project looks at solving these issues by proposing with several advantages will be given for the blind people than the existing system to identify the distance to object and relative velocity of the object along the path of the blind people. According to that, a voice will be produced to mention the distance to the object and relative velocity of the object along the path of the blind people by this device. According to that, blind can understand actual distance to the object and whether the object is coming towards to him or going away from him.
Augmented Reality Based Advertising System for Modern Home Items
(Uva Wellassa University of Sri Lanka, 2013) Wickramasinghe, A.A.M.W.
The main purpose behind this project concept is take full advantage of the internet and mobile world as a powerful marketing channel for profitable and interactive shopping. The system is capable for the control of the shopping experience is placed firmly in the consumer's hand with the advanced technological experience with Augmented Reality concept. AR is a generic term for an interactive 3D environment that blends with the physical reality, usually through a webcam or in this case a phone camera. The augmented reality and advertising is absolutely amazing so the customer interaction towards the business can be uplift easily with the proposed system. Augmented Reality (AR) is a upcoming advanced technology which can creatively used for the digital marketing. So it is very important for practicing these world newest technologies even within our countries if we want to stable in the global business process. And also Augmented Reality is still under the researcher so it is very interested to run a research under this type of field. The basic goals of the research can identified as augmented reality for usable business model. System accepts an image which contains the marker and recognized it. Finally system render the 3D model related to the particular given marker on the marker image. The users can experience the virtual 3D object interactively with the real environment. Therefore it allows users to visualize how a certain concept would look like in their home even before buying it. Augmented Reality, 3D rendering, Android platform were used to implement the system. The system consists of several modules including user interface module, intermediate module etc... The system performance varies based on the app launching device capabilities. This application couldn't launch with the use of emulator cause to the high performance capabilities that the application consumed during the running process. And also Java, XML used for the system implementation. The project is very innovative and interesting and it fully helps to uplift the the customer satisfaction and the efficiency of business by using e-commerce with proficient IT knowledge.
Automated Ceramic Tile Defect Recognition System
(Uva Wellassa University of Sri Lanka, 2013) Rangana, G.D.S.
In recent years, the real estate industry gets rapid development, and at the same time, the world ceramic tile production and consumption also get rapid development. Compared with the natural stone, polished porcelain floor tiles have many advantages of high strength, light weight, aging-resistance, corrosion resistance, and color consistency. As well as other industries quality assurance is one of the major tasks in ceramic tile industry. All ceramic tile Industries try to give competitive tiles without defects. In ceramic tiles there are different types of faults such as holes, dirty spots, color bleeding, pattern mismatch etc. Identification of these faults is very important for the ceramic tile industries. Automated Visual Inspection Systems are becoming increasingly popular due to low cost maintenance and high accuracy. Ceramic tile factories, for example, are very much interested in these sorts of systems. This system is used to surface defect recognition in automated way. Product line is used to carrying tiles to destination. Image is captured of each tile while tiles are flowing on the product line. After, captured image is supplied to the C# program. C# program analysis the image and compare with the original tile image. After, output signal is sent to circuit to stop process for defect tiles. Microcontroller is responsible for stop the motor of product line for defect tiles. Also system consists of data handling part for keep records of tiles. This proposed model will allow ceramic tile companies to perform quality control inspection without costly measuring tools or error-prone inspection by humans. Moreover, factories have to install and apply Flatness Control Machine to measure the flatness curvature of ceramic tiles. This machine keeps the ceramic tiles in fixed position to investigate the upper surface only.
Automatic Highway Ticketing System
(Uva Wellassa University of Sri Lanka, 2013) Chenthuri, J.
Southern Expressway is SriLankas first E class highway which reduces the time taken to travel from Colombo to Galle drastically. Even though several surveillance cameras are installed at the interchanges of the expressway, currently there is no way to automatically catalog the vehicles entering or leaving these interchanges. Thus in this work i utilized the image data acquired from these cameras to automatically log the journey information of all the vehicles travelling via the expressway, such that this information could be used for statistical analysis, fraudulent detection and managerial decision making. The project mainly consists of two components Common payment interface and intelligent highway port. I developed a scalable, distributed and real-time system which is partially invariant to illumination and rotation. I made several improvements to the existing methods as they failed due to the low resolution images, and incorporated intelligent algorithms, image processing techniques and domain knowledge to improve the accuracy.
Automatic Plant Leaves Recognition System
(Uva Wellassa University of Sri Lanka, 2013) Deliyakanthi, Y.S.D.
Today most of the people are interested in botany and students and researchers are engaging in botany. When studying botany, leaf recognition is an essential part. The different shapes and different nerve structures of the leaves are considered to classify the leaves. This research is developed to recognize a plant leaf by image. In the system highly considered about the nerve structure of the leaf. Basic Image processing Techniques, Artificial Neural Networks were used to implement the system. Identify a leaf using an image is a challenge. There are lots of different leaves around the world. It is not easy to identify all those leaves by single system. This research was developed to identify sample of 5types of leaves, Jack, Guava, Coral jasmine, lemon, Asoka. System accepts an image (one leaf with white background) and identifies the shape and nerve structure of the leaf. If it is pre identified leaf of the system it will give the details of that particular tree. Otherwise it will give a message to the admin to classify the leaf. Digital Image Processing and Artificial Neural Networks were used to implement the system. DIP module was used to extract the pictorial information from the source image, ANN module was used to recognize the nerve structure of the leaf image. Image processing module was implemented using C# programming language with the support of EmguCV (C# framework for OpenCV). .forge framework was used to implement the ANN module. All those modules and Graphical user interface (implement using C#) were integrated and form a single complete system that performed the process.
Automatic Surveillance System
(Uva Wellassa University of Sri Lanka, 2013) Ahilan, S.
As human we are automatically dealing with human behaviour. There are many industries in the world which require a system to observe human behaviour. Mainly for this purpose the resources required would be very high. But if an industry can maintain an automated system to maintain the surveillance then the resources can be used efficiently. When the complexity of the human activities got increased it causes to many frauds. Places where things such as gold, diamond are preserved and historical museums, bank lockers, and weapon reserves can be thrift by thieves easily. So it is very useful if I can maintain a surveillance system in those places to identify the activities and movements of an object. Then I tried to implement an "Automatic Surveillance system". Though some developed countries are using these kinds of systems, to the best of my knowledge nowhere in Sri Lanka there exists an Automatic Surveillance system. So my target is to minimize the human involvement, thus reducing the probability of occurrence of human errors and purposeful sabotage. That means authority can check the activities and trace when the objects moved. In my project I have used a video camera to capture the movement and then it transfers to the PC using wires. Then the motion detection software detects the movement and then an alarm is activated. The user can control the direction of the video camera through the software to follow and keep track of him. This is the process in brief that I intend to implement in my project.
Automatic Traffic Analysis Based Traffic Controlling System
(Uva Wellassa University of Sri Lanka, 2013) Krishnananthan, V.
Increasing transportation facilities raise the importance of an efficient traffic controlling system in modern world according to the changing traffic pattern. The most contemporary Traffic Controlling Systems are having enough technologies to control the traffic with highest efficiency without having a facility of taking the data regarding traffic details. The project attempts at solving these issues by proposing a solution which includes a traffic controlling mechanism. The approach to research is used to carry out a thorough research that works to identify the appropriate technology which influences detecting vehicles to increase efficiency, reliability & accuracy of System. The most suitable technology identified through the research is then used to design the solution. Furthermore, various issues, perspectives and theories that were identified through the research were used to model, design and eventually develop the IT solution.
Bone Crack Detector Based on X-Ray Using Fuzzy Logic and Neural Network
(Uva Wellassa University of Sri Lanka, 2013) Abesinghe, K.A.W.P.
Large numbers of X-Ray images are analyze by doctors in the hospitals day today. Based on that images doctors are going to predict what the problem in that bonds is such as bone crack, damage etc. These kinds of manual inspection of X-Ray consume a lot of time and the process itself is monotonous made during the inspection. The fractures of the bones will only occur during serious incident. This study discusses the development of a system which can differentiate the fractured bone from the non-fractured bone. The images of X-Ray differ in same bone also in different angle. The grown person the shape of the image of the bone is same from person to person. Also the shapes of the bones are same in healthy persons. Shapes of the bones are same in the grown person and baby. The continuous bones image is healthy one and fractions are classified first into three group which are Simple, Wedge and Complex. This study will focus on part of bone in the skeleton in our human body fracture detection. Basic Image processing Techniques, Fuzzy Logic and Neural Network were used to implement the system. System accepts the image and detects the crack of the long bone. This system is detecting the crack of femur bone and fingers. Image processing use to extract the bone from image, fuzzy logic use to detect crack of image form the features of bone and Neural Network verify that bone contains crack or not. System was implemented using C#.net programming language with the support of emguCV(C# wrapper for OpenCV) and Aforge.net. All those libraries and Graphical user interface were integrated and form a single complete system that performed the process. Separate the rough lines from the smooth lines to detect the crack of the image. This research showed that the Fuzzy logic with Digital Image Processing techniques, able to identify the crack of the bone using X Ray image with acceptable accuracy level.
Cheque Transaction Using Android Application
(Uva Wellassa University of Sri Lanka, 2013) Illamaran, B.
Nowadays in every banking industry using computer based applications to increase the quality of work and time efficiency. In order to improve the before mentioned areas, accuracy, reliability, security and speed of works has to be increased, if the particular bank has to win the competitive banking. In other words, dependency on human effort has to be reduced or new technologies have to be implemented. There are many banks in the world which handle large number of transactions at the same time. The banks use different data handling mechanism to manage all the cheque related data. The security and time are the major issues to be considered at such situations. As a solution, computerized technology has taken a very important part. To increase the efficiency of cheque transaction and managing mechanism with the existing system a dedicated system is needed to be implemented. So my system has done to fulfill all those expectations. This system is basically its handle cheque transaction with image processing, android application and GSM technology rather than the currently existing system. This System is a complete standalone application that is built to cheque transaction activities. This system is capable of sending image of cheque to server from android application. One of the special features of this system is able to process the information of the cheque. The system is able to send message to cheque owner's mobile and will do transaction based on his or her reply. The entire system is created using the Java, Android and GSM technologies. The cheque upload is done using the Android application with PHP. There is an interface which is done using Java, between server and client. The interface also handles image processing which is done using Nodejs and Tesseract OCR library. The SMS part is done using GSM technology with java. The database is created using Wamp server.
Criminal Identification System using fingerprint Plus Speaker Recognition
(Uva Wellassa University of Sri Lanka, 2013) Sundaralingam, T.
Generally biometric attributes are used in the criminal identification systems. Speaker recognition is referred as the identification of a person by the characteristics of the voice. This technique assists to verify the identity of an individual. Fingerprint authentication means the automated method for verifying matches between two human fingerprints. This research is based on speaker recognition and fingerprint recognition in criminal identification. Signal processing front end is the common technique used for speaker recognition and minutia extraction is used for fingerprint recognition. In this research both of the techniques have used for the speaker and fingerprint recognitions. The research is coded in MATLAB and SQL server ODBC connection is used for the database connection and the technology is neural network. Both recognitions contain the training and testing phase separately .Training phase is for recording voices and fingerprints pre stage. And in the testing phase the newly found voice clip or fingerprint is stored and compared with the previous output. Increased number of voice clips and fingerprints increases the efficiency of the system.
Detection and identification of Common Murmurs in Heart Using Artificial Neural Network
(Uva Wellassa University of Sri Lanka, 2013) Rajaguru, R.W.M.N.D.
According to the World Health Organization Heart diseases continue to remain within four major no communicable diseases in the world. So it is important to have an attention to every one's heart themselves. Today medical professionals are the one who uses stethoscope to check someone's heart. So It's important to implement a system a person can use a stethoscope at home and check whether his/her heart is healthier or not. It is a time and money saving effort. Professionals can use these systems to confirm their decision and they can make this a s a change for the old manual method. Recently, many research efforts have been carried out to apply artificial intelligence (AI) to auscultation based method for rigorous detection/classification of heart murmurs but accuracy rates are not always high. Most of the proposed systems have been using ECG or Electronic Stethoscope to implement to the system. In this project I tried to get the input through an Acoustic Stethoscope and implement the system. To implement the hardware Acoustic tubing was used and heart pulse is amplified and fed into computer as a digital signal. Recording was done there and clearly recorded sound was used to reduce noise. Then signal was used to feature extraction and neural network training. MATLAB language was used to create User Interfaces, segmentation and neural network. Time spectrum was analyzed for each heart sound .Supervised learning applied to Neural network using feed-forward algorithm.
Developing a 3D Game with Artificial Intelligence
(Uva Wellassa University of Sri Lanka, 2013) Priyadarshani, S.M.A.
Elements of AI used in computer games have come a long way. Probably the most often implemented form of AI, combat oriented AI can be seen in almost every shooter since the mid-nineties. Designers of first person shooters depend on combat AI to make their games playable, let along in interesting. With the development of game industry, more and more games started to focus on the AI aspect of game design instead of just graphics. Artificial Intelligence in games is usually used for creating player's opponents. Whether the player tries to sneak past them or go for a straight shoot, combat AI must decide how to react to the player in a very short time. From a games perspective, one key problem is the creation of AI driven agents that can interact with the player and be adaptive so as to create a great interactive gaming experience. A few examples of AI challenges in games include driving a car in a racing game, path finding on a map, planning behavior of non-player characters in a role-playing a game, resource gathering in a real time strategy game etc. However, computer games offer a great variety of other challenges including problems in graphics, sound, networking, player rating & match making, interface design, narrative generation, game world design, scripting etc. This research is aimed at building a 3D game for windows platform using Artificial Intelligence techniques. So, this first person shooter game is aiming to develop game with AI enemies who support animations, sounds, obstacle avoidance, target following, dynamic waypoint and path finding on a map. When playing the game, scores are given to each hunted enemy and displays high scores for player. Result of this project is a 3D game with Artificial Intelligence which makes game playable, let alone interesting and challengeable with a creative environment.
Drowsy Driver Detection System
(Uva Wellassa University of Sri Lanka, 2013) Niroshana, V.G.C.
Many road accidents happened due to driver's lack of attention while driving. It can happen due to many reasons. One of the major reasons is driver drowsiness. This paper describes a drowsy driver detection system which implements using c# language and Emgucv wrapper. The system consists of a web camera which continuously provides the video signal of driver face area. System detects drive's face, detects eyes, calculates the state of the eyes and if it detects a drowsy situation, plays an alarm to alert the driver. Haar like feature technique is used to detect the driver's face on frames. In this case there should be only one human face on fames. Haar classifier trained using equal number of positive and negative samples that are scaled to the same size. Positive samples contained various human faces which have open eyes. System uses face width and height to detect possible eye region to feed as input to detect state of the eyes. System uses two different methods to identify the state of the eyes. As the first method system uses change of intensity values around eye region because of that system can easily identify the eyebrow and upper lid. If the distance between eye brow and upper eye lid is greater than the system threshold value then eyes of the driver's are in close state otherwise eyes of the driver's are in open state. System uses support vector machine classifier to confirm the results of the first method to get a better accuracy in calculating state of eyes. Support vector machine trained using local binary pattern data of images. System extracts local binary pattern features from detected eye region and feeds them to support vector machine. Special algorithm designed to calculate the final state of the eyes using both results of haar classifier and support vector machine and identify the drowsy situation.
E-Learning Tool for Blind (ELTB)
(Uva Wellassa University of Sri Lanka, 2013) Vasiharan, P.
As there is a high demand in acquiring a well scoped E-learning tool for blind, a standardized as well as computerized system is preferred. This document describes the development of a standalone "E-learning tool for blind" (ELTB). This will be helpful for the students who are blind. Though it can be used by all the educational institution, my target is to implement this system at " Nuffield School for Deaf and Blind in Kaithady, Jaffna" as a trial. Therefore it focuses the need of blind school in such an E-learning tool. The system can be expanded for better performances in coming years with the help of the feedback I get while the system is being used. The document will further analyze the present situation and the changes that will be encountered when the system is installed and will also deal with the impact of this system on the targeted groups of users and the processes followed. The entire system is created using the Microsoft.Net and Sql server technologies. The tool is done using the Visual studio 2010 with C# language. This application will be developed using C#. This system is a standalone application so I created that system using latest technology called WPF. The text formats (Exam Questions, Subject Notes, Tutorials) which were used by the application stored using XML format. In database I used stored procedures for the better performance of the system. This E-learning tool contains five major components. Those are Subject Notes, Tutorials, Exams, Events and Reports. There I had three privilege levels those are super admin, admin and user. Super admin can be a principal of the school and admin can be a teacher. Here we used Google API called Google TTS for the conversion of text into speech. This needs high speed internet to connect with the API and to translate the text into speech. For the Speech to Text mechanism I created grammar file and got some help from the Microsoft Speech to Text API. For the creation of reports we used crystal reports. Those reports are very helpful for the teachers to filter the marks of the students and get to know the performance of them very easily.
E-Shopping Cart
(Uva Wellassa University of Sri Lanka, 2013) Chatumali, R.G.T.
Basically online shopping is a web based application.The main aim of online shopping is to improve the service of customers and venders. It maintains the details of the customer,payments,product receipts,addition of new customers,products and also updating,deletion.It also stores the details of invoices generated by customer and payments made by them with all Payments details like credit card. The primary features of online shopping are high accuracy, design flexibility and easy availability. It uses database tables representing entities and relationships between entities.
Emotion Detection from Digital Image
(Uva Wellassa University of Sri Lanka, 2013) Mary, G.P.
Emotion detection from digital image is presented in this research. The research is composed of four main stages, Face detection, eye detection, lip detection and emotion detection. First it takes an image, then by skin color segmentation it detects human skin color, then it detect human face. Then it separates the eyes and lip from the face. Then it draws bazier curve for eyes and lips. Then it compares the curve which that is stored in the data base. Then the emotion detection stage, the algorithm is adopted to recognize emotion from extracted features. It is shown by experiment results that the proposed method can detect emotion well. Although the technology for emotion recognition is important one which demanded in various fields, it still remains as the unsolved problem. Detecting emotion of human can be achieved by using facial image, voice, body shape, and etc. Among them, the facial image is most frequently source to detect emotion. While different ethnic groups have different levels of melanin and pigmentation, the range of colors that human facial skin takes on is clearly a subspace of the total color space. With the assumption of a typical photographic scenario, it would be clearly wise to take advantage of face color correlations to limit our face search to areas of an input image that have at least the correct color components. In pursuing this goal, we looked at three color spaces that have been reported to be useful in the literature, RGB and HSI spaces, as well as YCrCb. Segmenting an image is the process by which a computer attempts to separate objects within an image from the background as well as from other objects. The segmentation rules are the rules that will determine the formation of regions. The segmentation rules are based on analyzing the color and edge properties of a region. And accessibility problem also can be solved with this kind of the system; system can access the system anywhere, anytime he wants with maximum security. Interfaces are maximum simplifies therefore the user can interact with the system effectively.
Fault Signature Detection System
(Uva Wellassa University of Sri Lanka, 2013) Subasinghe, S.A.H.P.
In the current age the integration of image stitching algorithms, face detection algorithms could be used in many situations. And also the signature detection is used in many situations for diagnosis the correctness of the signatures. The research addresses to the problems with the fault signature detection and the correctness of the face detection as well using an easy way. This includes a fast algorithm of face detection for the frontal faces using haar cascade classifier. Throughout this paper the steps taken for the development of an innovative low-cost interactive fault signature detection system. Throughout this paper the steps taken for the development of an innovative low-cost interactive, efficient and less ambiguous fault signature detection system. Fault signature detection system created using JavaCV and Image Processing. Image Processing module was implemented using Java programming language with the support of JavaCV(Java wrapper for OpenCV).All those modules and Graphical user interface (implement using Java Swing) were integrated and form a single complete system that performed the process. There are many advantages of using this proposed system. Firstly the time effectiveness of this system is much lower than existing solutions. Secondly the ambiguous counting are reduce using this system. Thirdly Objectives of this system are successfully implemented with basic functionality of the image processing. Important parts of the system development were done successfully.Image Processing Module exhibit marginally accuracy when detecting the faces, System is basically detect the frontal faces of the students. Background light condition has limitation. This research showed that the Digital Image Processing techniques, able to get the stitch image of the class room, identify the number of faces and signatures with acceptable accuracy level.
Graphical User Interface Builder for Clanguage
(Uva Wellassa University of Sri Lanka, 2013) Jayaratne, T.V.P.D.
C language was originally developed in 1972 by Dennis Ritchie at Bell Laboratories. It is a programming language that derived from earlier language B and B's earlier ancestors BCPL and CPL. In a very short period of time C language became more powerful and flexible. The UNIX operating system was originally written in assembly language and then it was almost immediately re-written in C. During the rest of the 1970's, many colleges and universities were adapted C language because of its close ties to UNIX and the availability of C compilers. Though the c language found lasting use in applications, there is no specifically made RAD tool (only available one is "Glade" but it not specifically made for C language) that enable quick & easy development of user interfaces. Therefore most C programmers use GTK widgets to design graphical user interfaces using GTK widgets. This dissertation discusses about a RAD tool that provide programmers quick & easy development of user interfaces. Main objectives of this RAD tool is to provide a user friendly interface to design the graphical user interfaces, facility to preview the designed user interfaces and generate a C file to perform enhancement to the source code using code block or any other supported tool. First step of the procedure was to design an interface for the GUI builder. The interface was designed using java with eclipse environment. Then create an Editing area for design the graphical user interfaces to create windows forms, Insert buttons, labels, menus...etc with drag and drop facility. Next step was to read the attributes of the components when they are in their final positioning and then use file handling to create the C file using properties of above attributes. And compile the C file using GCC compiler and show the preview. Then as the final output, provide the GTK+ source code. This dissertation showed that implemented RAD tool will enable quick & easy development of user interfaces for C language.
Human Ear Recognition System
(Uva Wellassa University of Sri Lanka, 2013) Jayathilaka, W.M.S.D.
Human Ear Recognition is one of the challenging research areas nowadays. This research proposes a method to recognize human by using ear biometrics. Most of the other biometrics (Face, voice, and gait) change with age. When face was bearded and with emotions (smiling, crying, etc), it changes the basic shape and the appearance of the face. But in case of ear, the shape and appearance is fixed. Basically this research was focused to develop suitable algorithm to compare two ear images. This algorithm is based on Euclidian Distance. Before comparison there are several image processing activities to process ear image in order to eliminate unnecessary parts. They are Rotation, Cropping, Canny edge detection and rescale. This algorithm is based on image processing techniques. Java Programming language is used to implement this system. Java Advanced Imaging API is used to implement the Rotation, Cropping and Displaying components. Graphical User Interfaces are implemented using Java Swing. IIT Delhi Ear Database is used to test the system. These ear images are taken in several angles from same environmental conditions. So affect of the lightning is eliminated. This system archives 90% accuracy for this IIT Delhi Ear Database. But the accuracy level may change for other ear images taken from different environmental conditions. This System can be used for Criminal Investigation purposes. Police department, Custom department and other government agencies can use this system as a tool for person identification.

Browse

Browsing Abstracts of Undergraduates Dissertations-UWU by Subject "Computer Science and Technology"

Results Per Page

Sort Options