Concerns in Implementing Biometric Technology

Concerns in Implementing Biometric Technology Though this seems to be an advantage, the integration of this system into the existing system is tedious. Some of the major concerns in implementing biometric technology are as follows, The system relies on complex data processing algorithms which consumes considerable amount of time. Lack of manufacturing and integration of special purpose hardware in the existing system. Adoption of biometric technology in the day-to-day life is slow. A new approach that is gaining attention in the field of biometrics is referred as behavioral biometrics, also referred as behaviometrics. The behaviometrics concentrate on analysis the behavior of the user while interacting with the computer and try to authenticate him. The hardware Mouse capable of monitor the movements of the user and analyzing them to extract a signature, which is unique for every individuals [4]–[6]. Generally there two kinds of authentication mode available in the mouse dynamics, Static authentication Dynamic authentication The main strength of mouse dynamics biometric technology is in its ability to continuously monitor the legitimate and illegitimate users based on their session usage of a computer system. This is referred to as continuous authentication. Continuous authentication, or identity confirmation based on mouse dynamics, is very useful for continuous monitoring applications such as intrusion detection [5]–[8]. II. RELATED WORK Extensive research has been made in the field of utilising the oe of the computer input devices, Mouse, towards the development of user interface design structure [10]. Only in the recent times, the mouse dynamics is further improvised as behaviour biometric technology. The previous attempt ware made to study the user’s identity based on the mouse gesture analysis . Initially, the number of participants for this prgramme is around 48[12].The system is focused on both static and dynamic mode of authentication, but later the system exclusively tried to develop the continuous authentication because for static authentication where in the need of special purpose design of GUI and usage of certain predefined form of signature. Gamboa et al conducted similar experiments to learn the user’s movements while playing a memory game. They are 50 participants involved in the experiment. A sequential forward selection technique based on the greedy algorithm was simply used to find the best single feature later add one feature at a time to the feature vector. Gamboa et al[5] proved that increase in the movements (interactions), the more accurate the identification process would be. But, we cant use this approach to the static authentication type becau se Gamboa et al[5] reported that the memory game took 10-15 min in average. The main issues with these studies are the minimum amount of mouse movements required to authenticate an user was improbable. This method holds well for user reauthentication or continuous authentication but failed in static authentication. So, further work has to be done in the field of Mouse gesture dynamics: a behavior biometric [18], [19]. Our work is to identify the user based on their handwriting patterns. There are considerable amount of research work was made in the field of identifying the user based on his handwriting. The entire work process has been divided into two processes: signature verification and user identification. The pilot experiment where the 50 ample users are allowed to sign and their signature is later used to identify them. The participants are requested to draw eight different gesture and each of them twenty times. The same eight gestures are used throughout the entire process and the users are advised to draw the strokes in a single stroke. By studying pilot experiment meticulously, we can perceive following facts which play crucial role in our work and they are as follows. The average gesture size drawn was made up of 64 data points in a single stroke. Some participant tends to sign faster as they time goes and this cause departure from their normal behavior. The raw data contained noises that must be filtered before processing. The users were advised to be as consistent with the variability in shape and size. These variations were clearly a major source of inconsistency. In our paper, we provide security against shoulder surfing by toggling between the visibilities of the signature and also we provide additional security features like anonymous password feature. III. PROPOSED SYSTEM Based on the facts, we obtained from pilot experiment, we divided our entire work into following modules. Input gesture and sample modules Gesture processing Extraction and acquisition of data points Anonymous Password feature A. Input gesture and Sample modules The input gesture creation module and sample module is simple drawing screen that used to ask the participant to freely draw a set of predefined gestures. The main purpose of this module is to make the participant experienced with the system and to draw them in his own way which is to replicate them later on. So, the gestures are not bound to any specified language and they do not necessarily have a meaning. The input gesture creation and sample module helps the user in two different ways. First, it moves the input drawing to the center of the area. Though the shifting of the drawn gesture is done, the data points are collected without saving these changes. Second, the module moves the gesture spacing to achieve a size of 64 data points. These 64 data points were based on the pilot experiment. As mentioned earlier, we were able to determine the average size of drawing the predefined set of gestures in one stroke. B. Gesture Processing Once, the data is collected how these signatures are modified for further use. What are the steps involved in the process of converting the user signature into their corresponding data points are well briefed in this section. The signature collected from the drawing area consists of three main components, the horizontal coordinate (x-axis), vertical coordinate (y-axis), and the elapsed time in milliseconds at each pixel. Each gesture replication for a given gesture can be identified as the sequence of data points and each of them is represented by a triple consisting of the X-coordinate, Y-coordinate, and elapsed time, respectively. For example, the jth replication of a gesture G can be represented as a sequence Gj = {, , }, where n is referred to as the gesture size (GS) and each where (1≠¤ i ≠¤ n) is a data point. C. Extraction and acquisition of datapoints The extraction and acquisition of data points module involves three main components, namely, data acquisition, data preparation, and data storage and authentication. 1) Data Acquisition: This module presents the gestures, which was created initially by the user in the input gesture creation module, and displays them to the user to replicate. The module records the user’s drawing while he interact with the computer. This module essentially records the signature in three components, horizontal coordinates denoted by xij, vertical coordinates denoted by yij, and the elapsed time in milliseconds starting from the origin of the gesture tij, as explained in the input gesture module. For each user, the application creates individual folder containing all the replication of different gestures. Each gesture must be replicated a specific number of times (eg., 20 times). The user has to wait for minimum 3 s between each replication which is to prevent the user from drawing the gesture too fast. We believed that the wait time and mouse release will force the users back to his normal speed and behavior each time they replicate the gesture. 2) Data Preprocessing: This module is to process the collected data points in such a way it reduces to noise in it. The user’s signature may be shakened or jagged during drawing. They may lead to inconsistencies in the process of data point collections. There are two kinds of normalization techniques which should be applid first before reducing the noise patterns. The first is center normalization which shifts the gesture to the center of the drawing area. The idea behind this tranisition is that the user may tend to draw his signature at any corner of the drawing area so we need to process the signature from any any part of the area. So, it is advisable to move all the gestures to the center of drawing area. The second is size normalization which alters the size of the gesture so that the final size is equal to the size of the template gesture in order to compare the two gestures later. If the size of gesture is larger than the template size then k means algorithm is used to reduce its size. The k means algorithm forms 64 clusters of data points initially, take the centroids of each cluster as the datapoints. To remove the outliners and noise in each replication, data smoothing techniques are introduced. The user cant draw same signature without changing its size and shape under multiple occasions. So, the data smoothing removes the variations in the signature. We use the standard weighted least-squares regression (WLSR) method to smooth the data and Peirce’s criterion [21] to eliminate the outliers. 3) Data Storage and authentication: The collected data points are further stored in the database for each use. The database is capable of storing all the replication of gestures of the user which he entered during the input gesture and sample module. When the user entered the signature during the authentication time, all the replication gesture would be compared. is one of the imminent disaster in these modern technical world. Information extortion occurs when an attacker took the password and other authentication information from the user forcibly. Neither the traditional text-based password system nor biometric systems provide easy way-out of this. No matter the password is a text, fingerprint or iris movements it can be taken by force.