Character recognition using wavelet compression

The objective of this project is to build a character recognition system, which is able to recognize printed and handwritten character from A to Z. the typical optical character recognition systems, regardless the character´s nature, are based mainly on three stages, preprocessing, features extraction, and discrimination. Each stage has its own problems and effects on the system efficiency which is the time consuming and the recognition errors. In order to avoid these difficulties this project presents new construction of character recognition tool using the technique similar to that is used in image compression such as wavelet compression or JPEG compression. Wavelet compression is chosen as the technique implemented for this project. Wavelet compression technique extracted the important coefficient from the images. The Euclidean distance between the coefficient of the test images and training images is computed. Character is considered recognized if the Euclidean distance calculated is smaller than the Global threshold value of 258. This character recognition system also has 18.81% of false rejection rate and 21.88% for false acceptance rate.