The purpose of this project was to implement character recognition in mATLAB and to use this to break CAPTCHA's.A CAPTCHA can be an image or audio file included on a webpage that attempts to verify that a human is indeed filling out the form. Some image CAPTCHA’s may include different letters or numbers that go through a series of levels to add noise, waves, and warped letters. This approach makes it extremely difficult for a computer from running a basic Optical Recognition Model to decipher the text. To figure out methods to solve CAPTCHA’s we need to figure out what the problem with current methods are. This is really a two-part problem. One, how do we properly segment a distorted and warped image with lots of background noise. And two, how can we take this segmented image and create an intelligent character recognition algorithm to decipher the CAPTCHA. By developing methods to hack CAPTCHA’s we can in fact improve the current methods for doing Optical Character Recognition and segmentation for problems such as digitizing all scanned books while at the same time improving computer security against bots filling out web forms.