This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2016 conference. An efficient security system for data base management from illegal access V. D Ambeth Kumar, D. Elangovan Dept. of Computer Science and Engineering, Panimalar Engineering College, Chennai, India S. Lokesh Kumar, N. Eswaran Dept. of Computer Science and Engineering, Panimalar Engineering College, Chennai, India Abstract— The current scenario of scientific knowledge has seen a zoom in the improvement of various technologies. Attributable to this development, the importance level of protecting data has increased rapidly. Security to information defends the access to data from unauthorized users. The likelihood of knowledge insecurity arises to each giant and additionally tiny organization. Data Breach will have an effect on the consistency and reputation of the organization. At the moment to ensure security two major ways are followed particularly Numeric and Pattern. At the start Numeric manner was introduced to produce security which has variety locks, encrypted messages and revolving credit. Later came within the pattern technology that is far a lot of economical than Numeric’s. However to stress a lot of information preservation, we've to maneuver on to more developments. We have proposed a brand new system that's a lot of economical and effectual than the existing systems. Index Terms— Numeric’s, Patterns, Data Breach I. INTRODUCTION The effect of the modernized improvement in the technology has enlarged the urgency to implement high protection to the data. At Initial Stage Smartcard and Password were ideal for authentication [1].It provides simple instatement and exercise of user authentication. The attack of the malicious software tools on numeric password makes it more vulnerable. The risk of being forged in this type of authentication is more. The prevailing numerical schemes are also replaced by graphical schemes. But the user is more comfortable with the use of combination of text and graphical schemes [8]. This approach is implemented by a two step authentication method. In first step the numeric password is accepted from the user and in step two the user undergoes graphical password verification. In graphical representation the images are correctly selected as registered. Access to the database is granted if the two step verification is successful otherwise admittance is denied. On the other hand, Biometrics is automatic ways of recognizing someone supported a physiological or behavioral characteristic which tends to be unique that supports high level safety to data [9]. So to adapt to a new powerful authentication, biometric authentication methods are used. The past of life science includes the identification of individuals by distinctive body options, scars or a grouping of alternative physiological criteria, such like height, eye color and complexion. The current technologies include retinal scanning, measurement geometry, voice recording and facial recording [2]. Biometric technique is currently turning into the inspiration of a large array of extremely secure identification and private verification. Because the level of security breach and dealing scam increases the requirement for well secure identification and private verification. Recent world events had cause an increase interest in security which will impel life science into majority use. Areas of future use contain net transactions, workstation and network access, phone transactions. There have differing kinds of biometrics: Some square measure recent or others square measure latest technology. The foremost recognized biometric technologies square measure process, retinal scanning, hand pure mathematics, signature verification, voice recognition, biometric authentication and face recognition. Biometrics is suitable for user authentication and cannot be sensible to authenticate computer systems. The capture and transfer of biometrics should be conservative. Hence adopting a new scheme is important which comprises the existing methods (Numeric, pattern). Within the projected theme, the user will simply and with efficiency login system. II. RELATED WORKS Peter Auer and Zakria Hussain[3] planned a brand new technique for Implicit Feedback in Content-Based Image Retrieval exploitation pin read, that describes Pin read, a content-based image retrieval system that exploits implicit connexion feedback throughout an enquiry session. Pin read contains many novel ways that infer the intent of the user. From connexion feedback, like eye movements or clicks, and visual options of pictures Pin read learns a similarity metric between pictures that depend upon these interests of the user. It conjointly contains Pin read to the content-based image retrieval system PicSOM, so as to use it to world image databases. In[4] Vicent Alabau et al., enforced The naive Thomas Bayes model, generalizations and applications, that may be a terribly easy classification technique describes pattern options are assumed to be class-conditional freelance. this is often the alleged naive Thomas Bayes or independence assumption. In spite of being a robust, phantasmagoric assumption, the naive Thomas Bayes model typically provides smart results at low value in terms of model quality. This 850 978-1-4673-9338-6/16/$31.00 c 2016 IEEE