Abstract: Location is a primary indicator of context and forms the core basis of several context-aware applications. Most common way of getting location information is to use specialized hardware like GPS. However, GPS is expensive and is available only on high-end phones restricting its use to a smaller population in developing countries. Further, GPS also consumes a lot of battery power during its operation, thereby making it infeasible to run for longer durations with limited mobile phone battery. An alternative to GPS-based localization is GSM-based localization that is more suitable for developing countries due to much lower power consumption and ability to run even on low-end phones. Currently available, network-operator independent, GSM-based solutions require building perceptual map of cell towers in a city using war-driving.
In this paper, we present a novel low cost GSM-based solution based on Cell Broadcast (CBS) Messages. Location accuracy in our approach does not depend on building extensive cell ID database, typically built using war-driving. We present empirical studies (performed in the sub-city of Dwarka, New Delhi, India) comparing location accuracy of our approach with other GSM-based localization scheme that uses one of the most extensive open source database of cell IDs. We also compare power consumption of our proposed solution with GPS-based localization leading to energy-accuracy tradeoff that can be further exploited for a hybrid solution.
Abstract: Nowadays the amount of traffic flowing across the internet is increasing on a daily basis. This presents increasing risks to various entities on the internet right from threats like worms to distributed denial-of-service (DDoS) attacks. The problem of detecting these anomalies across the internet then becomes an important challenge. Anomaly detection on a large scale internet service provider (ISP) level has not been approached in a systematic manner. We propose CASCADE (Cloud Based Anomaly Detection) a distributed design for privacy-preserving sharing of information where it is impossible for ISPs to find out private and sensitive information about each other. As part of our analysis we first examine a well accepted algorithm (PCA) which detects anomalies in origin-destination flows across multiple ISPs. We then consider non-linear approaches using laplacian-eigenmaps. We measured distributed implementations of PCA and Laplacian Eigenmap algorithms across some performance parameters like messages exchanged between two ISPs, time taken for computation and detection rates. The experimental results are promising and we outline scope for further improvement.
Abstract: In this project we have worked on building “CVDmagic application” to identify people who have high risk to cardio-vascular diseases without doing any “laboratory test” for rural Indian settings. We have worked on building CVDmagic application such that it does not require any laboratory test. Lack of desired healthcare resources, Non-efficient uses of available resources, non-adherence of village people to treatment compliance etc. are other generic health care issues present in rural-Indian health care settings. We have tried to tackle all these issues, while designing our CVDmagic application.
Abstract: Newborn swapping and abduction is a global problem and traditional approaches such as ID bracelets and footprinting do not provide the required level of security. This paper introduces the concept of using face recognition for identifying newborns and presents an automatic face recognition algorithm. The proposed multiresolution algorithm extracts Speeded up robust features and local binary patterns from different levels of Gaussian pyramid. The feature descriptors obtained at each Gaussian level are combined using weighted sum rule. On a newborn face database of 34 babies, the proposed algorithm yields rank-1 identification accuracy of 86.9%.
Abstract: Here are various modes of image acquisition. It can be visible spectrum imaging, acoustic imaging, x-ray imaging, infrared imaging, and thermal imaging. Each of these has it’s own advantages and limitations. In context to face recognition, it has been observed that the near infrared (NIR) imaging is less prone to illumination variation which is a fundamental challenge in the domain application. On the other hand, visible spectrum imaging provides local and global cues important for face recognition. In this research, we are leveraging the robustness of NIR imaging for illumination variation along with the discriminatory information provided by the visible spectrum. The information extracted from these two imaging domains can be combined at different fusion levels. We have developed a hardware setup for capturing both the visible and NIR images simultaneously. Currently, we are developing novel algorithms for extracting unique and discriminatory information from the multimodal images and combining them to improve the performance of face recognition.
Abstract: This paper presents an efficient algorithm for matching sketches with digital face images. The algorithm extracts discriminating information present in local facial regions at different levels of granularity. Both sketches and digital images are decomposed into multi-resolution pyramid to conserve high frequency information which forms the discriminating facial patterns. Extended uniform circular local binary pattern based descriptors use these patterns to form a unique signature of the face image. Further, for matching, a genetic optimization based approach is proposed to find the optimum weights corresponding to each facial region. The information obtained from different levels of Laplacian pyramid are combined to improve the identification accuracy. Experimental results on sketch-digital image pairs from the CUHK and IIIT-D databases show that the proposed algorithm can provide better identification performance compared to existing algorithms.
Abstract: There are many results on the possibilities of privacy leak in location-based services, and many remedies have also been suggested. In this paper, we look at this problem from an altogether different perspective - claiming that privacy leaks are most likely inevitable for a lot of remedial mechanisms, and therefore it is necessary to provide means to a user to detect such leaks in advance. We specifically target the mechanisms using location obfuscation (aka. blurring) for providing location privacy for continuous queries. We give a general set of definitions to measure leak of location privacy. We also give a framework to compute the leak for a wide range of location obfuscation strategies. We then evaluate a few common strategies both analytically and experimentally and use the results to propose provably efficient algorithms for them.
Abstract: Monitoring and controlling the energy consumption of indoor appliances are crucial not only to save money but also to better utilize the energy sources. We present a low power and low cost energy monitoring system that monitors and controls the energy consumption of appliances remotely using mobile phones. DTMF signals, generated for each key tone on mobile phone, are used to send control information. A mobile phone attached to the system receives the DTMF signals and passes it to the controlling system over its audio jack. The control information can also be sent to the system via SMS. Any readings from the monitoring system are sent to the attached mobile using Bluetooth and the same will be sent back to the caller through SMS. Our proposed system also focusing on minimizing power consumption during sensing by providing fine grained control of system components.
Abstract: Due to increasing use of Short Message Service (SMS) over mobile phones in Developing Countries, there has been a burst in number of spam SMSes. Even after regulatory solutions like DND registry in India, SMS spam seems to be increasing and causing a lot of annoyance to users.There is lack of technological and user focused SMS spam filtering solution. Content-based machine learning approaches were effective in filtering email spams. Unlike email spam filtering, SMS spam filtering can be largely influenced by the presence of regional words, abbreviations and idioms and extensive feature engineering is required before using machine learning techniques for classification. We also found that the perception of SMS were different for men and women through a survey. In this paper, we have tested the feasibility of applying Bayesian learning and SVM based machine learning techniques which were reported to be most effective in email spam filtering on a India centric dataset. We have also developed a user focused mobile-based system SMSAssassin that can filter SMS spam messages based on bayesian learning and sender blacklisting mechanism. Since the spam SMS keywords and patterns keep on changing, SMSAssassin uses crowd sourcing to keep itself updated. Using a dataset that we are collecting from users in the real-world, we evaluated our approaches and found some interesting results. As a future work, we are developing a SVM based spam filtering system which will make use of cloud for better spam filtering accuracy thus making it a promising Mobile-Cloud application.
Abstract: The aim of the project is sneeze detection on Maemo based phones such as Nokia N900. It is motivated by a larger Healthcare project which combines the results of this module and other captured/interpreted data to arrive at a conclusive disease of flu. The technique involves recording an audio sample using Nokia N900’s recorder application and using this audio file as an input to the program. The implementation consists of taking FFT of the audio signal and using amplitude along with root mean square to determine if the audio was a sneeze or voice. Work is still undergoing to distinguish between sneeze and shout..
Abstract: We propose Stegobot, a new generation botnet that communicates over probabilistically unobservable communication channels. It is designed to spread via social malware attacks and steal information from its victims. Unlike conventional botnets, Stegobot tra ffic does not introduce new communication endpoints between bots. Instead, it is based on a model of covert communication over a social-network overlay - bot to botmaster communication takes place along the edges of a social network. Further, bots use image steganography to hide the presence of communication within image sharing behavior of user interaction. We show that it is possible to design such a botnet even with a less than optimal routing mechanism such as restricted flooding. We analyzed a real-world dataset of image sharing between members of an online social network. Analysis of Stegobot's network throughput indicates that stealthy as it is, it is also functionally powerful - capable of channeling fair quantities of sensitive data from its victims to the botmaster at tens of megabytes every month.
Abstract: Defect tracking systems facilitate software maintenance activities such as issue reporting, triaging, developer collaboration and bug fixing. The quality of bug report submitted to defect tracking systems is a topic that has attracted a lot of research attention recently. Previous studies show that the quality of information present in a bug report influences its resolution time and has impact on the productivity of the development team .Bug reports contain certain mandatory and standard fields like Product Name, Component Name, Version Number, Platform and Operating System. Studies shows that bug reporters often provide incorrect values for such mandatory and standard fields. We investigate the phenomenon of incorrect information in standard fields and throw light on the process-data quality issue specifically pertaining to incorrect assignment of Components in bug reports. We perform empirical analysis on publicly available bug repository of Eclipse project and report correlations between variables and features extracted from bug report data and the phenomenon of incorrect assignment of Component Name. We use the findings of our statistical analysis to develop a predictive model to support practitioners in identifying incorrect component assignments in real-time.
Abstract: Open source software projects are characterized as self organizing and dynamic in which volunteers around the world primarily driven by self-motivation (and not necessarily monetary compensation) contribute and collaborate to a soft- ware product. In contrast to close source or proprietary software, the organizational structure and task allocation in an open source project setting is unstructured. Soft- ware project managers perform risk, threat and vulnerability analysis to gain insights into the organizational structure for de-risking or risk mitigation. For example, it is important for a project manager to have an understanding of critical employees, core team, subject matter experts, sub-groups, leaders and communication bridges. Software repositories such as defect tracking systems, versioning systems and mailing lists contains a wealth of valuable information that can be mined for solving practically useful software engineering tasks. In this paper, we present a systematic approach to mine defect tracking system for risk, threat and vulnerability analysis in a software project. We derive a collaboration network from a defect tracking system and apply social network analysis techniques to investigate the derived network for the purpose of risk and vulnerability analysis. We perform empirical analysis on bug report data of Mozilla Firefox project and present the results of our analysis. We demonstrate how important information pertaining to risk and vulnerability can be uncovered using network analysis techniques from static record keeping software archive such as the bug tracking system.
Abstract: Skin detection is a challenging problem having a wide range of computer vision applications. For example, it can be used as an added classifier in face detection. However, obtaining an accurate skin color model is a difficult task in adverse illumination, different camera settings and diverse ethnicity.Static skin color models have fixed thresholds and rely on image preprocessing to bring skin pixel values within these boundaries. Therefore, the performance, in terms of accuracy and computation time, degrades drastically in real world fast video applications. In this paper, we propose an adaptive skin color model to reduce the false accept cases of Adaboost face detector. Since the face color distribution model is regularly updated using previous Adaboost responses, the algorithm is more effective to real world environmental co-variates. Further, the adaptive nature of the skin classifier does not significantly affect the computation time; hence the algorithm has several applications in surveillance.
Abstract: As we all know that the need for a connection between the users is becoming an indispensable need in the growing world of technology, therefore we aimed to achieve a similar sort of a goal using the concepts of Distributed Systems using Android mobile phones. The project aims to target any small organisation like IIIT Delhi or software companies which are all covered via wireless networks. One such useful example can be foreseen in our own institute when some student needs an e-book and he/she doesn’t know who has that particular book. One way to get the e-book is ask everyone one by one and then get it through email or using pen drive. Instead of doing all this hard work why shouldn't we use technology to address our problem? Here comes the use of this project, just start the application on your android phone and broadcast a message requesting the ebook. Anyone having the book can reply to the person and send the book on the mobile itself.
Our project targeted a real world application by connecting Android mobile phone users through a Wi-Fi network and facilitating various actions from File sharing to message broadcasting. This application enable users to chat, send/receive queries and share files where all of them will be connected by a wireless network. The system will be totally decentralised and will be based on P2P model
Abstract: Automata Plus is a set of graphical tools to learn basic concepts of Automata Theory. The application is designed for experimenting with Theory of Computation topics including nondeterministic and deterministic finite automata , several types of grammars , Regular Expressions and Pumping Lemma. It allows one to convert one form to other thereby verifying proofs as well as learn the concepts on their own. AutomataPlus has GUI based , user friendly highly interactive interface .The interface has been developed in PyQt4 , a standard GUI toolkit for Windows. We designed the application in a modular way so as to enable the use of individual modules as standalone packages , which can be used as an external library to other applications or even to be extended later .
Abstract: Our aim was to predict the probability of a team to win a One Day International / Twenty-20 cricket match while the game is in progress and keep updating the probability of winning after every ball. We have successfully designed a working live model which can actually predict the outcome of the match at any particular instant and is updated after every ball. We had coded two models simultaneously :
Abstract: Traffic congestion is becoming a serious problem in modern cities. Encouraging more private-vehicle drivers to use public transportation is one of the most effective and economical ways to reduce the ever increasing congestion problem on the streets. Thus, to make public transport services more attractive and competitive, providing travellers with individual travel advice for journeys becomes crucial.However, with the massive and complex network of a modern city, finding one or several suitable route(s) according to user preferences from one place to another is not a simple task. From a passenger point of view, identifying routes is very complicated simply because of the enormous size of public transport systems.
Our algorithm implements optimized versions of graph algorithm keeping in mind scalability and running times for large graphs and thus returns route to users in minimum time. Routes are improved such that it returns a path with minimum stops and interchanges (least time and lowest fare path). Caching is implemented to reduce the algorithm usage for same query, and keeping scalability in mind. Data was collected from websites of DTC and Delhi Metro, due to incomplete and unambiguous data, data cleaning was required. Later version of the product implements the precise datasets provided by DIMTS. Entire service is integrated with a SMS service in which user can make a query using SMS and hence can get the route details while he / she is in motion.
Abstract: Random numbers have wide applications and are used in cryptography, casinos, lotteries, scientific applications, etc. Thus there is a need for highly reliable and secure random number generators. Most of the random no. generators available are pseudo random number generators (PRNGs). They are algorithms that use mathematical formulae or simply pre-calculated tables to produce sequences of numbers that appear random. But actually the sequences are pre-determined and depend entirely on a seed value. Once the seed is known the entire sequence can be generated again. On the other hand True random number generators measure some physical phenomena that are expected to be random and introduce it into a computer which produces random numbers from the collected data. They have wide applications in cryptography which need a highly secure and non-deterministic RNG. In this project we have generated data by capturing the background noise with the help of a microphone and by tracking mouse movement. Thus the data generated is entirely random and each bit is independent of the previous bit. We captured sound at various frequencies ranging from 8 KHz to 200 KHz. We performed tests by using the NIST test suite which is a world-wide accepted standard for testing and analyzing the generator. The test suite results showed that sound with frequency ranging from 16000 Hz to 24000 Hz passed all the tests and hence the generator can be considered suitable for generating random numbers and can also be used in cryptographic applications.
Abstract: Is uploading pictures onto Social Networking sites second nature to you? Do you spend considerable amount of time a day on OSNs leafing through pictures that others put up? Do you store your important data such as passwords and PIN numbers on your computer? Well, then you are the perfect target for DreamWalker!
A botnet is a network of computers on the internet that have been set up (without the knowledge of their owners) to forward transmissions (data/spams/viruses etc) to other computers on the internet. Dreamwalker is a special kind of botnet – the zombies here (the unsuspecting computers on the internet that are a part of the botnet) are the users of a popular OSN – Facebook. The botnet thrives by hiding data from the zombie’s computer (passwords/”hidden” data in a folder) into the pictures that the user uploads on Facebook (Steganography). When the next zombie “views” the pictures, the information in them is extracted and uploaded again whenever he/she uploads a picture. And thus the chain continues till it reaches the Botmaster (the one who controls all the zombies). Ironically enough, the Botmaster too is a mere profile on Facebook!
The network lies upon the network of Facebook users (Overlay Network) and is almost impossible to detect (Imperceptible) as it creates no extra traffic. Given the statistics (number of Facebook users today – 600 million, and the number of pictures uploaded by users on this New Year alone – 750 million) this is a potentially fatal botnet. Our project aims to prove that the concept is indeed feasible on Facebook...
Abstract: Most smart phones have an inbuilt FM transmitter.It is provided to transmit a song being played on the phone to for example the stereo system in the car.The FM transmitter normally available in the smart phones has a range from 88 Mhz to 109 Mhz.It is capable of transmitting any song being played (or a beeping tone otherwise) in a range of 4-8 m.This FM transmitter can be exploited in a number of ways.For eg. it can be used as a medium of communication ,specially in areas where the usual cell network does not exist like rural areas.With this aim, we tried to access the FM transmitter of Nokia 900 and successfully transmitted a signal at any desired frequency.This has opened scope for creating applications on Nokia 900 which can exploit the transmitter.We also created an application called Object Finder which aims to make searching real world objects like wallet, pendrive etc as simple as searching a file on a computer.We propose that receivers be attached to objects which can then be searched using the transmitter in the phone.The application provides an interface to the user to register the objects with their frequencies and then search the object on a click of a button.
Abstract: The Study delves into the domains of Multi-Party Computation(MPC) and Game Theory.The intersection of both these fi elds is the interaction amongst mutually distrusting parties. MPC Model is based on computation of a function based on the inputs provided by parties.The Game Theoretic Model gauranties some payoff based on the joint actions of all players and individual contribution,so each player tries to maximize its payoff. Secret sharing on the other hand refers to method of distributing information aka 'Secret' amongst a group of participants, each of which is allocated a share of the secret. The secret can be reconstructed only when a sufficient number of shares are combined together; individual shares are of no use on their own.Collection of Shares can independently lead to reconstruction of the Secret, this scheme is also known shamir secret sharing scheme. The study model takes up the essence of rational secret sharing introduced by Halpern and Teague(STOC2004).It tries to emulate the classical problems in domains of Secret Sharing and Multi-Party Computation by conducting an experiment amongst undergraduate students at IIIT Delhi.The Study Model is inspired from simple Game theory postulates which state that every player in a game is rational.The study model takes a paradigm shift from Loyal and Adversary participants to Rational players. A Secret Sharing Game is conducted in which participants share propositional logic statements in rst order logic to deduce the truth values of variables which is the secret,The game lasts for a total of 5 rounds and statements are shared through a trust party.The Player to deduce truth values of most variables correctly wins the game.The game shows that players work towards attaining an unstable nash equilibria where their set of strategies is strongly vary by unrational choices made by other players, whereas some favorable changes in the rules of the game and guidance of the trust party improves the rational behavior of players substantially thus the game attains a correlated equilibria.
Abstract: Most image search engines use text based tags for searching. Although text is easier to index and search, the accuracy of search decreases as an image might not have the correct tags around it. Thus, it produces a lot of garbage in results. In Content Based Image Retrieval (CBIR), the images are searched according their similarity in colour, texture, shape and other visual features and hence promise to be more accurate. In this project we have implemented some CBIR algorithms based on colour, texture, shape and a fusion approach. We have divided the process into two parts namely, indexing and matching and tried to optimize both for each approach. We have also created a test database and evaluated the accuracy of the different algorithms.
Abstract: LivePNR , a service to provide easy hassle-free access to users , to the Indian Railways PNR retrieval system. The application developed on a cloud-based model would make it easier for the users to keep track of the changes to their PNR status , and thereby eliminating the need to check repeatedly for it . This will relieve the users from all the hassle of opening a web-browser , navigating to the Indian Railways Website and retrieving PNR status manually again and again for changes. Hosting the service using a cloud-based model , would reduce the costs to a great extent of running such a service , and even some revenue could possibly be generated using small ads in the updates being provided to the user. The service can also come to an aid to provide other Indian Railways related services where once could register to be receive constant updates about the running status of trains.
Abstract: The past few years have seen an explosion of location sharing applications which allow individuals to share their location with a prescribed set of people. All the existing applications like Google latitude, Yahoo’s Fire Eagle rely on GPS technology or Wi-Fi units to track user's location which is being provided in the high end mobile phones only. As a result, these applications have failed to grab a foothold in the developing nations where the usual masses do not have access to such sophisticated technologies. We, here, propose a location sharing application which is designed keeping in mind these constraints, to be able to reach out to the masses of a developing nation. We have used CBS messages to track user’s location and like many other popular applications, we use Facebook to allow him to connect to his friends and see their locations. SMS is used as a communication medium to update user’s location on the server as opposed to the much costlier Internet/GPS. Hence, the service provides novelty on three fronts, namely Platform Independent, Subscriber Independent and Indian Specific application. The Web application is hosted on the MUC server and the mobile app is also already up and working. We are planning to roll out the application soon to study the privacy preferences of the people on the basis of usage of the application by a constrained set of users.
Abstract: We have worked on a Java-based mobile client for Android phones called Moca. Moca is used to collect health related data using a phone and send the data back to an open source backend database called OpenMRS. Currently the interface between Moca and OpenMRS happens through an intermediate layer called Moca Dispatch Server (MDS). Thus, MDS is a software layer, written in Python, that collects data coming from the phone and pushes it into the database. In Indian context, most of the programmable phones support Symbian OS instead of Android OS. In this project we developed a client similar to Moca for Symbian OS while keeping the MDS layer (interfacing with OpenMRS) intact.
Developing countries like India face the double burden of infectious and chronic diseases. Given the high cost of treating chronic diseases, it’s imperative that health systems in the country are re-oriented towards prevention. Sana is the centerpiece of a learning system that is designed not only to improve clinical outcomes, but the health delivery process itself. Thus the dire need of such a system drove us to develop upon the same to improve the system all together.
Abstract: Student feedback for a course plays a very vital role in shaping the overall course structure and teaching modalities for the course. In our institute, students give their feedback at the end of the course on paper-based forms. However it is only befitting for an IT centric institute to make this process of taking feedback online. Keeping this in mind, we have developed a web application, Online course Feedback system (OCFS).
The application is built using the .Net framework. It has been designed in such a way that it requires minimal of human intervention. For each course the admin takes time in the order of minutes filling out the details. Our major concern during the development of this application was to ensure the anonymity of the student. To ensure that, we do not store the students’ responses against their identity at any level of the process. The faculty too never interacts with the web application and directly receives the result of the feedback in his/her inbox. The summary of the responses in every way preserves the anonymity of every student. The Application has been deployed on the IIIT-D server and has performed without any glitches in the test phase. This application will be used in the coming semesters to take the feedback from the students.
Abstract: The phenomenal rise of online social networks over the past decade has seen millions of users putting up a wealth of personal information on the World Wide web. This research study aimed to utilize the publically available information on one such online social network Facebook to gain useful insights about a subject and thereby do context aware social phishing attacks. Three different parameters namely mutual friends, network and gender were considered while structuring the attacks and based on their level of susceptibility of falling for such attacks, users were categorized into three distinct classes. Overall, it was established that incorporating the social context significantly increased the yield of a phishing attack. Subjects were also found to be duped when the attack emancipated from a mutual friend or the same network. Males were found to be prone to phishing attacks which had their origins in the opposite gender.
Abstract: The Auto rickshaw fares in New Delhi tend to have little relation with meter fares as a very small percentage of auto rickshaw drivers use the meters. Again the accuracy of the metering is also doubtful, with several news stories highlighting various techniques auto drivers use to swindle passengers. Hence our application RightFARE, which gives a very accurate idea about the auto/taxi fares for New Delhi.
RightFARE is an sms based fare prediction system for New Delhi that can give an accurate idea about the auto/taxi meter fare between two locations The basic problem that we undertook was to find out an accurate meter fare between two locations for an auto rickshaw which by just changing the calculations could easily be extended to taxis. RightFare is a combination of fare calculation and user-reported data that gives a complete picture of expected auto fares from one point to another. If a user sends an SMS to our predefined number with the origin and destination, the RightFare server connect with the Google Maps API and Wikimapia and our own databse of geo-locations to retrieve the shortest route and return the distance. This distance by simple calculations can then be converted to fare amount which is then given to the user by sms.
Abstract: Peer Review System (PRS) is a free Web-based program that allows instructors to incorporate frequent reviewing of reports/assignments into their courses, regardless of class size, with limited instructional resources. This will be like a conference management system (ex. Easychair) where the Instructor will be the PC Chair, students in the class will be the PC members, as well as authors/submitters. The project has been implemented in C# and ASP.NET using MySQL database. There are three consoles – instructor, admin and student. This project has been deployed on the IIIT server after rigorous system testing and is being used for reviewing in TW course in this semester. Lot of work has been done on GUI and documentation (SRS, Design, Testing plan, project management plan) as well. PRS is a unique program that is structured so that you will learn to critique academic writing and apply learned techniques to your fellow student’s (PEER’S) assignments. PRS should be user-friendly, ‘quick to learn’, reliable and independent software for the above purpose. After the students submit their work, the PRS program guides them through a tutorial on peer reviewing of that particular assignment. The PRS writing assignments allow you to explore assigned topics in greater depth and with more critical thinking.
Abstract: In a world where video surveillance has become the need of the hour, its authentication still poses to be a major issue. Be it a surveillance video or any other video recording, its genuineness remains to be the concern of many. These videos, especially when used for legal issues, have to be ensured to be tamper- free. With plenty of tools available which can change or alter the contents of a video by a mere click, this problem becomes more severe and of utmost importance.
The proposed algorithm uses change detection to perform video authentication. We follow a prediction based approach to predict the value of a pixel at a particular location and then compare this value with the original pixel value. The predicted value is calculated using the weighted sum of the pixels at the same location in the previous frames. If the difference between these values is greater than a pre-specified threshold then that particular location is considered to be tampered. The similar approach is followed for a set of specific pixels for a frame and by checking how many locations are tampered, we decide whether the frame is tampered or not.
Abstract: The system is aimed towards a recruitment agency namely Lecan Solutions Pvt. Ltd. which is contacted by organizations to hire people. The system models the workflow of the recruitment agency which can have multiple organizations as clients. It starts with client organization providing the recruitment agency some specifications, and the system based on the specifications provided will look into the database of resumes and come up with some candidates who are suitable for the job. After this, the selected candidates will be called for further interviews by the organization and their status will be regularly updated by the system. Once all the interviews are over and the client organization is finally completed with the hiring, the system should generate an invoice.
© 2011 IIIT-Delhi. All Rights Reserved. |
|---|