FRACTAL RECOGNITION AND PATTERN CLASSIFIER BASED SPAM FILTERING IN EMAIL SERVICE Ms. S.Revathi 1, Mr. T. Prabahar Godwin James 2 1 Post Graduate Student, Department of Computer Applications, Sri Sairam Engineering College, Chennai, Tamilnadu. India. 2 Assistant Professor, Department of Computer Applications, Sri Sairam Engineering College, Chennai, Tamilnadu. India. Abstract: Efficient monitoring and preventing process requires allocation and policies to network administrator implies network security. The main communication link for individual and organizational conversation is an E-mail. Spam mails introduced in mail communication. Flooding of internet due to many copies of passed typical message depicts spam. Keyword filtering utilized Bayesian algorithm to get rid of spam emails. But, the detection performance of Bayesian algorithm based models is poor. Spot approach used in this paper to shoot out the mails; blocking the domain without referring the users E-mail ID, keyword based blocking by monitoring the subjects, password security by bio-metric. Moreover, this paper introduces two recognition techniques such as contour detection technique and Fractal detection / recognition technique using the image of the user is introduced. Keywords- Face recognition, Fractal code, Contour point, Pattern classifier, Spam Filter. 1. INTRODUCTION Electronic mail, a method that exchanges digital messages from a sender to one or more receiver abbreviated as e-mail. Gmail is the globally required web- based email provider with maximum active users worldwide such as 425 million users. Computer application that identifies and verifies the person from the digital image without human intervention is termed as facial recognition system [1-4]. Initially, image was stored in data base. Then, the comparison of selected facial features with the data base [ 5 ]. It can be used in security systems. A successful source of information to discover the persons is human face. Hence, the differentiation of persons in spite of facial similarity is performed by using recognition [6, 7]. Face recognition systems are categorized into three approaches namely, local, global and hybrid as follows: Local approaches: Identification of a person depends upon high discriminating parts and their geometrical relationship between them such as eyes, nose and mouth. Global approaches: The human face considered as whole object and included information in global approaches. The methods used in global approaches contains discrete cosine transform, Gabor wavelet and eigen faces Google provides open email service as Gmail. Gmail is an open email service provided by Google. Real time users access the protected webmail like Gmail through POP3 or IMAP4 protocols. Gmail contains search oriented interface and chat view. Hybrid approaches: Human face recognition performed by local and global features by using the principle imitation of human visual system. The fractal representation used to handle the local and global features to ensure the recognition rate and recognition time. The unsolicited and unwanted emails are detected by a program termed as spam filter [1]. Here, spam filter checks all incoming emails to your email accounts against mail filter rules. 2. EXISTING SYSTEM Information transmissions between users in different locations are done through email processes effectively. The validity of the delivery of message on the form of identification of is performed by Authentication. In existing system, the detection and recognition of face is used to identify human being for authentication Bayesian algorithm [7] is applied in traditional methods to obtain the rid of spam emails. DISADVANTAGES: The evolving nature of spam mails is not completely avoided by spam keyword filtering. Automatic email content examining and spamming is not possible. To filter out spam domains is not possible. Since a p p e a r a n c e b a s e d m e t h o d s are used for face recognition identifying twin people or people who met with accidents are difficult. 3. PROPOSED SYSTEM In proposed system, secure login Gmail; services and efficient spamming are considered for high effective authentication. Authentication based on Contour detection fractal 1 Page August 2015, Volume - 2, Issue - 4
detection/ recognition is performed on user s image introduced in this paper. Since fractal detection and recognition is a unique method to identify every human being. The authentication to service leads to effective fractal detection. The proposed system comprises pattern classifiers, such as Keywords and URL s for data check, tag construction and keyword identity, automatic reading of mails is the concepts used in this system. P-ISSN: 2347-4408 3.1 HUMAN FACE RECOGNITION BASED ON FRACTAL IMAGE CODING Human face recognition is an important area in the biometrics field [2]. One of the first steps of the face recognition is to detect and extract face from an image. Once face image has been detected, recognition can be carried out. Here we present fractal methods for face recognition. Initially, the face images of candidates identified by interconnection of pixels. The pixel s interdependence represented as fractal code in the form of chain of pixels. The mathematical p r i n c i p l e b e h i n d t h e a p p l i c a t i o n of fractal image codes for recognition is derived as follows: The image (x f ) with fractal parameters A and B is given by, x f = Ax f + B (1) The dissimilar fractal codes from the definition of fractal transformation for arbitrary image is given by, (T x = A x x f + x f (2) Fig.1 Proposed System Architecture Filtering of spam mails and keywords is carried out by proper maintaining of repository of pattern classifiers [6]. The proposed filtering process frequently removes the surplus mails in same domain and different mail id. The flaws in spam filtering overcome by the conceptions by spammed keyword [1]. Spot approach for spotting the mails has been introduced. If a person sending number of mails to a respective person with in a stipulated time then the mail will redirect to administrator for approval. This approach helps in a way like a spammer can send number of mails from the same mail ID or same domain. ADVANTAGES: Authentication by means of sequential processes such as contour detection fractal detection/recognition provided better results. Pattern classifier based spam filtering makes the proposed more thriving. Fig.2 Face recognition process The proposed fractal recognition implemented in various steps as follows: Normalization of the face image. Feature extraction using fractal coding of the normalized face image. Face localization is an uncomplicated form of face detection. If any two face feature points are detected Then that face can be located and normalized. Therefore, face localization can also be tackled via facial feature detection [2].The eyes, nose and mouth were identified using straight image processing techniques. Assume that the horizontal position of nose was also determined and an accurate locus 2 Page August 2015, Volume - 2, Issue - 4
for the nose tip is predicted. The estimation of human face involves the identification of the loci of feature points namely, eyes, nose and mouth [2]. Here we used geometrical normalization, where find some points in the face image (contour points) then find the orientation of points. Fig.3 contour points Algorithm1: Fractal Coding Input: Normalized image Output: Partition t h e image int o non-overlapping range blocks Ri using quad-tree partitioning method. Cover the image with sequence of overlapping domain blocks Dj. For each range blocks, find the domain block and corresponding transformation that will match the range block. Set the geometrical positions of the range blocks and matching domain blocks as well as matching transformation as fractal code of face image. P-ISSN: 2347-4408 problem for most of the users and we don t have proper solution in manipulating. The following key issues like, Email scanning before it s read by the users. Blocking t h e d o m a i n i r r e s p e c t i v e of t h e users email id. Keyword based blocking by subjects monitoring. A program used to detect the unwanted email and prevention of unwanted messages was spam filter [3]. Here spam filter checks all incoming emails to your email accounts against mail filter principles. The tags are outlined and HTML tags are automatic parsed [4]. The comparison between the tags and URL s was done by using the top down parsing. The email identified as spam mail for matched tags. The report describes spam emails and URL s prepared for future email rejections in private domain. The repository for received URL s and spam emails maintained by an administrator. There are number of pattern matching algorithms [6]. Here we used Brute Force algorithm for top down key word parsing. Algorithm 2: Brute-Force Pattern Matching Input: text Txt of size n and pattern Pt of size m Output: starting index of a Substring of Txt equal to Pt or -1 if no such substring exists for (i = 0; i< n ; i ++) j = 0; while (j <m && Txt[i + j] = = Pt[j]) j = j + 1; if ( j== m) return I; return -1; The brute-force pattern matching algorithm compares the pattern Pt with the text Txt for each possible shift of Pt relative to Txt, until either n a match is found. Brute-force pattern matching runs in time O (nm). 4. IMPLEMENTATION Fig.4 An illustration of domain and range blocks 3.2 PATTERN BASED SPAM FILTEING The problem in avoiding spam is an identification of patterns. Nowadays, spams are considered as one of the major technical 3 Page August 2015, Volume - 2, Issue - 4
Fig.5 Detect face and register The above figure has shown the detection of face and registration process. Fig.8 Spot checking for mails The above Fig.8 has shown spot checking for mails. Fig.6 Adding spam information Fig.6 has used to adding spam information in the database. Fig.9 Admin Approval/Reject phase Fig.9 has shown the admin login for approval or evaluate phase Fig.7 Detecting spam URLs Detecting in spam URLs has shown in Fig.7 4 Page August 2015, Volume - 2, Issue - 4
Fig.10 Evaluation phase The above Fig.10 has shown the performance evaluation phase. 5. CONCLUSION In this paper, a high effective authentication scheme for secure log in email services and effective spamming are taken in to consideration. The concepts in relation to pattern classifiers namely, URL s for checking data, construction of tag, keyword identity and automatic reading of mails used in this system. REFERENCES [1] H. Alkahtani, P. G. Stephen, and R. Goodwin, "A taxonomy of email SPAM filters," 2011. [2] R. Jafri and H. R. Arabnia, "A survey of face recognition techniques," journal of information processing systems, vol. 5, pp. 41-68, 2009. [3] H.-Y. Lam and D.-Y. Yeung, "A learning approach to spam detection based on social networks," Hong Kong University of Science and Technology, 2007. [4] W. Ma, D. Tran, and D. Sharma, "On Extendable Software Architecture for Spam Email Filtering," in I MECS, 2007, pp. 924-928. [5] A. Adler, "Vulnerabilities in biometric encryption systems," in Audio-and Video-Based Biometric Person Authentication, 2005, pp. 1100-1109. [6] A. A. Cárdenas and J. S. Baras, "Evaluation of classifiers: practical considerations for security applications," in AAAI Workshop on Evaluation Methods for Machine Learning, 2006, pp. 409-415. [7] D. Lowd and C. Meek, "Good Word Attacks on Statistical Spam Filters," in CEAS, 2005. 5 Page August 2015, Volume - 2, Issue - 4