The Captcha-Cracker started as a college project. The aim of this project is to allow a "layman" to select a captcha scheme based on a set of sample images. Currently, the project attempts to break 18 different types of 2d text captchas and a 3d captcha.

• Better recognition modules
• Interface for preprocessing and segmentation (in developer's corner)
• Is it possible to reduce dependence on font specific training set ?