Character Recognition Based on Euclidean Distance Calculation Over Predictive and Centroid Coordinates

Abstract

Now a day people are living with internet technology but those technologies brings many problems to the people through many hacking techniques. Image spam is the one among them. In the earlier stages, hackers used to annoy targeted victims with their fabricated text called spam text. Hackers are passing their bogus information on many ways such as advertising, spam emails, buttons, query distributions etc. From which spam emails are very specific to attack and they are filtered by text based filter. Then attackers nurtured their attacks on new way i.e., spreading spam mails by images. Those images are non related content to the concerned users on their corresponding mails or any web pages. Because of those spam images, text based filter couldn’t identify spam texts. On the basis of an image’s features, Attackers used to embed their spam text or mischief coded links into some of the attracted images. To identify spam contents from an image, security functions of a system must be able to recognize the characters imbedding on any images. This research paper is going to present views on image spam, Data mining approaches for dataset analysis, proposed optical character recognizer model and implementation of character recognition from images using Euclidean distance values.